Professional Documents
Culture Documents
Saaty and Vargas (2012) 2nd Ed - Models, Methods, Concepts & Applications of The Analytic Hierarchy Process
Saaty and Vargas (2012) 2nd Ed - Models, Methods, Concepts & Applications of The Analytic Hierarchy Process
Volume 175
Series Editor
Frederick S. Hillier
Stanford University, CA, USA
123
Thomas L. Saaty Luis G. Vargas
Katz Graduate School of Business and Katz Graduate School of Business and
College of Business Administration College of Business Administration
University of Pittsburgh University of Pittsburgh
Mervis Hall, Pittsburgh Mervis Hall, Pittsburgh
PA 15260 PA 15260
USA USA
ISSN 0884-8289
ISBN 978-1-4614-3596-9 e-ISBN 978-1-4614-3597-6
DOI 10.1007/978-1-4614-3597-6
Springer New York Heidelberg Dordrecht London
v
vi Preface
vii
viii Contents
5 Designing a Mousetrap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.2 Effectiveness Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.3 Attracting the Mouse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.4 The Trap Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.5 The Cost and Benefits of the Trap . . . . . . . . . . . . . . . . . . . . 75
5.6 A Marketing Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.7 The Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
19 Abortion and the States: How will the Supreme Court Rule
on the Upcoming Pennsylvania Abortion Issue? . . . . . . . . . . . . . . 281
19.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
19.2 Subcriteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
19.3 Alternatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
19.4 Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
19.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Chapter 1
How to Make a Decision
1.1 Introduction
9ieqa9 a9qvg9 is the Greek word for hierarchy meaning holy origin or holy rule.1 It is
the ordering of parts or elements of a whole from the highest to the lowest. A
hierarchy is the principle of control that secures the effective functioning of the
organization.2
‘‘You can’t compare apples and oranges,’’ so the saying goes. But is this true?
Consider a hungry person who likes both apples and oranges and is offered a
choice between a large, red, pungent, juicy looking Washington State apple and an
even larger, old and shriveled, pale colored orange with a soft spot. Which one is
that person more likely to choose? Let us reverse the situation and offer the same
person on the next day a small, deformed, unripe apple with a couple of worm
holes and a fresh colored navel orange from California. Which one is he or she
more likely to choose now?
We have learned through experience to identify properties and establish
selection criteria for apples and oranges and in fact we use that experience to make
tradeoffs among the properties and reach a decision. We choose the apple or
orange that yields, according to our preferences, the greater value across all the
various attributes.
The Analytic Hierarchy Process (AHP) is a basic approach to decision making.
It is designed to cope with both the rational and the intuitive to select the best from
a number of alternatives evaluated with respect to several criteria. In this process,
the decision maker carries out simple pairwise comparison judgments which are
then used to develop overall priorities for ranking the alternatives. The AHP both
allows for inconsistency in the judgments and provides a means to improve
consistency.
1
Encyclopedia Catholica.
2
The Great Soviet Encyclopedia, Moscow 1970.
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 1
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_1, Springer Science+Business Media New York 2012
2 1 How to Make a Decision
Perhaps the most creative task in making a decision is deciding what factors to
include in the hierarchic structure. When constructing hierarchies one must include
enough relevant detail to represent the problem as thoroughly as possible, but not
so thoroughly as to lose sensitivity to change in the elements. Considering the
environment surrounding the problem, identifying the issues or attributes that one
feels should contribute to the solution, and who are the participants associated with
the problem, are all important issues when constructing a hierarchy. Arranging the
goals, attributes, issues, and stakeholders in a hierarchy serves two purposes: It
provides an overall view of the complex relationships inherent in the situation and
in the judgment process, and it also allows the decision maker to assess whether he
or she is comparing issues of the same order of magnitude.
The elements being compared should be homogeneous. The hierarchy does not
need to be complete; that is, an element in a given level does not have to function
as a criterion for all the elements in the level below. Thus a hierarchy can be
divided into subhierarchies sharing only a common topmost element. Further, a
decision maker can insert or eliminate levels and elements as necessary to clarify
the task of setting priorities or to sharpen the focus on one or more parts of the
system. Elements that are of less immediate interest can be represented in general
terms at the higher level of the hierarchy and elements of critical importance to the
problem at hand can be developed in greater depth and specificity. The task of
setting priorities requires that the criteria, the subcriteria, the properties or features
of the alternatives be compared among themselves in relation to the elements of
the next higher level.
Finally, after judgments have been made on the impact of all the elements, and
priorities have been computed for the hierarchy as a whole, sometimes, and with
care, the less important elements can be dropped from further consideration
because of their relatively small impact on the overall objective.
1.3 Philosophy, Procedure and Practice of the AHP 3
Fig. 1.2 Structural difference between a linear and a non linear network
For completeness we mention that there are four axioms in the AHP. Briefly
and informally they are concerned with the reciprocal relation, comparison of
homogeneous elements, hierarchic and systems dependence and with expectations
about the validity of the rank and value of the outcome and their dependence on
the structure and its extension [7].
Cognitive psychologists have recognized for some time that there are two kinds of
comparisons that humans make: absolute and relative. In absolute comparisons,
alternatives are compared with a standard or baseline which exists in one’s
memory and has been developed through experience. In relative comparisons,
alternatives are compared in pairs according to a common attribute. The AHP has
been used with both types of comparisons to derive ratio scales of measurement.
We call such scales absolute and relative measurement scales. Relative mea-
surement wi, i = 1,…, n, of each of n elements is a ratio scale of values assigned
to that element and derived by comparing it in pairs with the others. In paired
comparisons two elements i and j are compared with respect to a property they
have in common. The smaller i is used as the unit and the larger j is estimated as a
multiple of that unit in the form (wi/wj)/1 where the ratio wi/wj is taken from a
fundamental scale of absolute values.
Absolute measurement (sometimes called scoring) is applied to rank the
alternatives in terms of either the criteria or the ratings (or intensities) of the
criteria; for example: excellent, very good, good, average, below average, poor,
1.4 Absolute and Relative Measurement and Structural Information 5
and very poor; or A, B, C, D, E, F, and G. After setting priorities for the criteria (or
subcriteria, if there are any), pairwise comparisons are also made between the
ratings themselves to set priorities for them under each criterion and dividing each
of their priorities by the largest rated intensity to get the ideal intensity. Finally,
alternatives are scored by checking off their respective ratings under each criterion
and summing these ratings for all the criteria. This produces a ratio scale score for
the alternative. The scores thus obtained of the alternatives can in the end be
normalized by dividing each one by their sum.
Absolute measurement has been used in a variety of applications. For example,
it has been used to rank cities in the United States according to nine criteria as
judged by six different people [13]. Another appropriate use for absolute mea-
surement is in school admissions as in Chap. 22 [14]. Most schools set their criteria
for admission independently of the performance of the current crop of students
seeking admission. The school’s priorities are then used to determine whether a
given student meets the standard set for qualification. Generally, candidates are
compared with previously set standard rather than with each other. In that case
absolute measurement should be used to determine which students meet prior
standards and qualify for admission.
The following two examples provide partial validation of the 1–9 scale used in
the pairwise comparisons of homogeneous elements.
Which drink is consumed more in the US?
A B C D E F G Estimated Actual
A:Coffee 1 9 5 2 1 1 1/2 0.177 0.18
B:Wine 1/9 1 1/3 1/9 1/9 1/9 1/9 0.019 0.01
C:Tea 1/5 3 1 1/3 1/4 1/3 1/9 0.042 0.04
D:Beer 1/2 9 3 1 1/2 1 1/3 0.116 0.12
E:Sodas 1 9 4 2 1 2 1/2 0.190 0.18
F:Milk 1 9 3 1 1/2 1 1/3 0.129 0.14
G:Water 2 9 9 3 2 3 1 0.327 0.33
C.R. = 0.022
1.5 The Fundamental Scale 7
A B C D E F G Estimated Actual
A:Steak 1 9 9 6 4 5 1 0.345 0.37
B:Potatoes 1/9 1 1 1/2 1/4 1/3 1/4 0.031 0.04
C:Apples 1/9 1 1 1/3 1/3 1/5 1/9 0.030 0.00
C:Soybeans 1/6 2 3 1 1/2 1 1/6 0.065 0.07
E:Whole wheat bread 1/4 4 3 2 1 3 1/3 0.124 0.11
F:Tasty cake 1/5 3 5 1 1/3 1 1/5 0.078 0.09
G:Fish 1 4 9 6 3 5 1 0.328 0.32
C.R. = 0.028
There is an infinite number of ways to derive the vector of priorities from the
matrix (aij). But emphasis on consistency leads to the eigenvalue formulation
Aw = nw. To see this, assume that the priorities w = (w1,…,wn) with respect to a
single criterion are known, such as the weights of stones, we can examine what we
8 1 How to Make a Decision
eij = 1 ? dij, dij [ - 1, and substituting in the expression for kmax. It is dij that
concerns us as the error component and its value |dij| \ 1 for an unbiased esti-
mator. The measure of inconsistency can be used to successively improve the
consistency of judgments. The consistency index of a matrix of comparisons is
given by C.I. = (kmax - n)/(n - 1). The consistency ratio (C.R.) is obtained by
comparing the C.I. with the appropriate one of the following set of numbers (See
Table 1.2) each of which is an average random consistency index derived from a
sample of randomly generated reciprocal matrices using the scale 1/9, 1/8,…, 1,…,
8, 9. If it is not less than 0.10, study the problem and revise the judgments.
The AHP includes a consistency index for an entire hierarchy. An inconsistency of
10 percent or less implies that the adjustment is small compared to the actual
values of the eigenvector entries. A proof that the number of elements should be
small to preserve consistency can be found in [6].
Perhaps the most creative and influential part of decision making is the structuring
of the decision as a hierarchy. The basic principle to follow in creating this
structure is always to see if one can answer the following question: ‘‘Can I
compare the elements on a lower level in terms of some or all of the elements on
the next higher level?’’
A useful way to proceed is to work down from the goal as far as one can and
then work up from the alternatives until the levels of the two processes are linked
in such a way as to make comparison possible. Here are some suggestions for an
elaborate design.
1. Identify overall goal. What are you trying to accomplish? What is the main
question?
2. Identify subgoals of overall goal. If relevant, identify time horizons that affect
the decision.
3. Identify criteria that must be satisfied in order to fulfill the subgoals of the
overall goal.
4. Identify subcriteria under each criterion. Note that criteria or subcriteria may
be specified in terms of ranges of values of parameters or in terms of verbal
intensities such as high, medium, low.
5. Identify actors involved.
6. Identify actor goals.
7. Identify actor policies.
10 1 How to Make a Decision
All science is descriptive not normative. It is based on the notion that knowledge is
incomplete. It uses language and mathematics to understand, describe and predict
events with the object of testing the accuracy of the theory. Events involve two
things: (1) controllable and uncontrollable conditions (e.g. laws) and (2) people or
1.10 Normative: Descriptive 11
In its simplest form, the AHP begins with the traditional concept of ordinal
ranking to stratify a hierarchy and advances further into numerical paired com-
parisons from which a ranking of the elements in each level is derived. By
imposing a multiplicative structure on the numbers (aijajk = aik), the reciprocal
condition is obtained. Thus, the AHP infers behavioral characteristics of judg-
ments (inconsistency and intransitivity) from its basic framework of paired com-
parisons. It begins by taking situations with a known underlying ratio scale and
hence known comparison ratios, and shows how its method of deriving a scale
uniquely through the eigenvector gives back the original scale. Then, through
perturbation the AHP shows that a derived scale should continue (through the
eigenvector) to approximate the original scale providing that there is high
consistency.
1.11 Rationality
1.12 Examples
the goal, and on the third (or bottom) level are the three candidate houses that are
to be evaluated in terms of the criteria on the second level. The definitions of the
factor and the pictorial representation of the hierarchy follow.
The factors important to the family are:
1. Size of House: Storage space; size of rooms; number of rooms; total area of
house.
2. Transportation: Convenience and proximity of bus service.
3. Neighborhood: Degree of traffic, security, taxes, physical condition of sur-
rounding buildings.
4. Age of House: Self-explanatory.
5. Yard Space: Includes front, back, and side space, and space shared with
neighbors.
6. Modern Facilities: Dishwasher, garbage disposal, air conditioning, alarm
system, and other such items.
7. General Condition: Extent to which repairs are needed; condition of walls,
carpet, drapes, wiring; cleanliness.
8. Financing: Availability of assumable mortgage, seller financing, or bank
financing.
The next step is comparative judgment. The elements on the second level are
arranged into a matrix and the family buying the house makes judgments about the
relative importance of the elements with respect to the overall goal, Satisfaction
with House.
The questions to ask when comparing two criteria are of the following kind: of
the two alternatives being compared, which is considered more important by the
family and how much more important is it with respect to family satisfaction with
the house, which is the overall goal?
The matrix of pairwise comparisons of the factors given by the home buyers in
this case is shown in Table 1.3, along with the resulting vector of priorities. The
judgments are entered using the Fundamental Scale, first verbally as indicated in
the scale and then associating the corresponding number. The vector of priorities is
the principal eigenvector of the matrix. This vector gives the relative priority of the
14 1 How to Make a Decision
factors measured on a ratio scale. That is, these priorities are unique to within
multiplication by a positive constant. However, if one ensures that they sum to one
they are always unique. In this case financing has the highest priority, with 33% of
the influence.
In Table 1.3, instead of naming the criteria, we use the number previously
associated with each.
Note for example that in comparing Size of House on the left with Size of
House on top, a value of equal is assigned. However, when comparing it with
Transportation it is strongly preferred and a 5 is entered in the (1, 2) or first row,
second column position. The reciprocal value 1/5 is automatically entered in the
(2, 1) position. Again when Size of House in the first row is compared with
General Condition in the seventh column, it is not preferred but is moderately
dominated by General Condition and a 1/3 value is entered in the (1, 7) position. A
3 is then automatically entered in the (7, 1) position. The consistency ration C.R. is
equal to 0.169 and one needs to explore the inconsistencies in the matrix with the
help of Expert Choice to locate the most inconsistent one and attempt to improve it
if there is flexibility in the judgment. Otherwise, one looks at the second most
inconsistent judgment and attempts to improve it and so on.
We now move to the pairwise comparisons of the houses on the bottom level,
comparing them pairwise with respect to how much better one is than the other in
satisfying each criterion on the second level. Thus there are eight 3 9 3 matrices
of judgments since there are eight elements on level two, and three houses to be
pairwise compared for each element. The matrices (Table 1.4) contain the judg-
ments of the family involved. In order to facilitate understanding of the judgments,
a brief description of the houses is given below.
House A: This house is the largest of them all. It is located in a good neigh-
borhood with little traffic and low taxes. Its yard space is comparably larger than
that of houses B and C. However, its general condition is not very good and it
needs cleaning and painting. Also, the financing is unsatisfactory because it would
have to be financed through a bank at a high interest.
House B: This house is a little smaller than House A and is not close to a bus
route. The neighborhood gives one the feeling of insecurity because of traffic
conditions. The yard space is fairly small and the house lacks the basic modern
1.12 Examples 15
facilities. On the other hand, its general condition is very good. Also an assumable
mortgage is obtainable, which means the financing is good with a rather low
interest rate. There are several copies of B in the neighborhood.
House C: House C is very small and has few modern facilities. The neigh-
borhood has high taxes, but is in good condition and seems secure. The yard space
16 1 How to Make a Decision
is bigger than that of House B, but is not comparable to House A’s spacious
surroundings. The general condition of the house is good, and it has a pretty carpet
and drapes. The financing is better than for A but not better than for B.
Table 1.4 gives the matrices of the houses and their local priorities with respect
to the elements on level two.
The next step is to synthesize the priorities. There are two ways of doing that.
One is the distributive mode. In order to establish the composite or global priorities
of the houses we lay out in a matrix (Table 1.5) the local priorities of the houses
with respect to each criterion and multiply each column of vectors by the priority
of the corresponding criterion and add across each row, which results in the
composite or global priority vector of the houses. The other way of synthesizing is
the ideal mode. Here the priorities of the houses for each criterion are first divided
by the largest value among them (Table 1.5). That alternative becomes the ideal
and receives a value of 1. One then multiplies by the priority of the corresponding
criterion and adds as before. House A is preferred if for example copies of B matter
and hence the distributed mode is used. In a large number of situations with 10
criteria and 3 alternatives, the two modes gave the same best choice 92% of the
time [7]. House B is the preferred house if the family wanted the best house
regardless of other houses and how many copies of it there are in the neighborhood
and hence the ideal mode is used.
Employees are evaluated for raises. The criteria are Dependability, Education,
Experience, and Quality, Each criterion is subdivided into intensities, standards, or
subcriteria as shown in Fig. 1.4. Priorities are set for the criteria by comparing
them in pairs, and these priorities are then given in a matrix. The intensities are
then pairwise compared according to priority with respect to their parent criterion
1.13 Absolute Measurement 17
(as in Table 1.6) and their priorities are divided by the largest intensity for each
criterion (second column of priorities in Fig. 1.4). Finally, each individual is rated
in Table 1.7 by assigning the intensity rating that applies to him or her under each
criterion. The scores of these subcriteria are weighted by the priority of that
criterion and summed to derive a total ratio scale score for the individual. This
approach can be used whenever it is possible to set priorities for intensities of
criteria, which is usually possible when sufficient experience with a given oper-
ation has been accumulated.
The City of Pittsburgh has become a leader in the world in organ transplantations.
Because there are more patients who need livers, hearts and kidneys than there are
18 1 How to Make a Decision
available organs, it has become essential to assign priorities to the patients. The priorities
shown in the figures are a result of several months of study by Alison R. Casciato and
John P. O’Keefe in coordination with doctors and research scientists at a local hospital.
Absolute measurement was used for this purpose and is shown in Fig. 1.5a–c. The
hierarchy of Fig. 1.5 consists of the goal, the major criteria and subcriteria after which
some of the subcriteria are further divided into yet smaller subcriteria or are divided into
intensities for rating the patient. Figure 1.5a–c give further subdivision into intensities
for those subcriteria in Fig. 1.5 that need to be further subdivide d into intensities. In
general, one would use the intensities to score a patient. When there is no intensity, either
the full value of the criterion is assigned, or a zero value otherwise. For example,
Criminal has 0.033 priority and that value is awarded to a patient with no criminal record.
A patient with a criminal record would receive a zero. The goal is divided into: emo-
tionally dependent and financially dependent patients: Both are divided into single,
married, and divorced with and without dependent and financially dependent patients.
Then each of them is further divided into: medical history (time on donor list, degree of
disability), physical history (degree of ability to endure rehabilitation, willingness to
cooperate, etc.), and social status (criminal record, volunteer work). The priorities are
indicated next to each factor and sum to one for each level. A patient is ranked according
to the intensities under each criterion. The higher the total score the better the opportunity
to receive a transplant.
In addition to the many illustrations given in this book, the AHP has been used in
the economics/management area in subjects including auditing, database selection,
design, architecture, finance, macro-economic forecasting, marketing (consumer
choice, product design and development, strategy), planning, portfolio selection,
facility location, forecasting, resource allocation (budget, energy, health, project),
sequential decisions, policy/strategy, transportation, water research, and perfor-
mance analysis. In political problems, the AHP is used in such areas as arms
control, conflicts and negotiation, political candidacy, security assessments, war
games, and world influence. For social concerns, it is applied in education,
1.14 Applications in Industry and Government 19
(a)
(b)
(c)
Bibliography
1. Dyer RF, Forman EH (1989) An analytic framework for marketing decisions: text and cases.
Prentice-Hall, Englewood Cliffs
2. Expert Choice Software, Expert Choice, Inc. (1994) 4922 Ellsworth Ave., Pittsburgh 15213
3. Golden BL, Harker PT, Wasil EA (1989) Applications of the analytic hierarchy process.
Springer, Berlin
4. Kinoshita E (1993) The AHP method and application. Sumisho Publishing Company, Tokyo
5. Saaty TL (1990) Decision making for leaders, RWS Publications, 4922 Ellsworth Ave.
Pittsburgh. First appeared 1982 Wadsworth, Belmont
6. Saaty TL (1990) The analytic hierarchy process, paperback edition, RWS Publications,
Pittsburgh. First appeared 1980, McGraw Hill, New York
7. Saaty TL (1994) Fundamentals of decision making and priority theory, RWS Publications,
4922 Ellworth Ave., Pittsburgh
8. Saaty TL, Alexander J (1989) Conflict resolution. Praeger, New York
9. Saaty TL, Kearns KP (1985) Analytical planning–the organization of systems, international
series in modern applied mathematics and computer science 7. Pergamon Press, Oxford
10. Saaty TL, Vargas LG (1982) The logic of priorities applications in business, energy, health,
transportation. Kluwer-Nijhoff Publishing, Boston
Bibliography 21
11. Saaty TL, Vargas LG (1991) Prediction, projection and forecasting. Kluwer Academic,
Boston
12. Saaty TL, Vargas LG (1993) A model of neural impulse firing and synthesis. J Math Psychol
37:200–219
13. Saaty TL (1986) Absolute and relative measurement with the AHP: the most livable cities in
the United States. Socio-Econ Planning Sci 20(6):327–331
14. Saaty TL, France JW, Valentine KR (1991) Modeling the graduate business school
admissions process. Socio-Econ Planning Sci 25(2):155–162
15. Saaty TL (1986) Axiomatic foundation of the analytic hierarchy process. Manage Sci
32(7):841–855
16. Saaty TL (1993) What is relative measurement? The ratio scale phantom. Math Comput
Model 17(4–5):1–12
17. Simon HA (1955) A behavioral model of rational choice. Quart J Econ 69:99–118
18. Tversky A, Slovic P, Kahneman D (1990) The causes of preference reversal. Am Econ Rev
80(1):204–215
Chapter 2
The Seven Pillars of the Analytic
Hierarchy Process
2.1 Introduction
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 23
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_2, Springer Science+Business Media New York 2012
24 2 The Seven Pillars of the Analytic Hierarchy Process
The seven pillars of the AHP are: (1) Ratio scales, proportionality, and
normalized ratio scales are central to the generation and synthesis of priorities,
whether in the AHP or in any multicriteria method that needs to integrate existing
ratio scale measurements with its own derived scales; in addition, ratio scales are the
only way to generalize a decision theory to the case of dependence and feedback
because ratio scales can be both multiplied, and added—when they belong to the
same scale such as a priority scale; when two judges arrive at two different ratio scales
for the same problem one needs to test the compatibility of their answers and accept
or reject their closeness. The AHP has a non-statistical index for doing this. Ratio
scales can also be used to make decisions within an even more general framework
involving several hierarchies for benefits, costs, opportunities and risks, and using a
common criterion such as economic to ensure commensurability; ratio scales are
essential in proportionate resource allocation as in linear programming, recently
generalized to deal with relative measurement for both the objective function and the
constraints obtaining a ratio scale solution vector form which it is possible to decide
on the relative values of the allocated resources; one can associate with each alter-
native a vector of benefits, costs, time of completion, etc., to determine the best
alternative subject to all these general concerns; (2) Reciprocal paired comparisons
are used to express judgments semantically automatically linking them to a
numerical fundamental scale of absolute numbers (derived from stimulus response
relations) from which the principal eigenvector of priorities is then derived; the
eigenvector shows the dominance of each element with respect to the other elements;
an element that does not have a particular property is automatically assigned the
value zero in the eigenvector without including it in the comparisons; dominance
along all possible paths is obtained by raising the matrix to powers and normalizing
the sum of the rows; inconsistency in judgment is allowed and a measure for it is
provided which can direct the decision maker in both improving judgment and
arriving at a better understanding of the problem; scientific procedures for giving less
than the full set of n(n - 1)/2 judgments in a matrix have been developed; using
interval judgments eventually leading to the use of optimization and statistical
procedures is a complex process which is often replaced by comparing ranges of
values of the criteria, performing sensitivity analysis, and relying on conditions for
the insensitivity of the eigenvector to perturbations in the judgments; the judgments
may be considered as random variables with probability distributions; the AHP has at
least three modes for arriving at a ranking of the alternatives: (a) Relative, which
ranks a few alternatives by comparing them in pairs and is particularly useful in new
and exploratory decisions, (b) Absolute, which rates an unlimited number of alter-
natives one at a time on intensity scales constructed separately for each covering
criterion and is particularly useful in decisions where there is considerable knowl-
edge to judge the relative importance of the intensities and develop priorities for
them; if desired, a few of the top rated alternatives can then be compared against each
other using the relative mode to obtain further refinement of the priorities; (c)
Benchmarking, which ranks alternatives by including a known alternative in the
group and comparing the other against it; (3) Sensitivity of the principal right
eigenvector to perturbation in judgments limits the number of elements in each set of
2.1 Introduction 25
comparisons to a few and requires that they be homogeneous; the left eigenvector is
only meaningful as reciprocal; due to the choice of a unit as one of the two elements in
each paired comparison to determine the relative dominance of the second element, it
is not possible to derive the principal left eigenvector directly from paired compar-
isons as the dominant element cannot be decomposed a priori; as a result, to ask for
how much less one element is than another we must take the reciprocal of what we get
by asking how much more the larger element is; (4) Homogeneity and clustering are
used to extend the fundamental scale gradually from cluster to adjacent cluster,
eventually enlarging the scale from 1–9 to 1–?; (5) Synthesis that can be extended
to dependence and feedback is applied to the derived ratio scales to create a
uni-dimensional ratio scale for representing the overall outcome. Synthesis of the
scales derived in the decision structure can only be made to yield correct outcomes on
known scales by additive weighting. It should be carefully noted that additive
weighting in a hierarchical structure leads to a multilinear form and hence is non-
linear. It is known that under very general conditions such multilinear forms are
dense in general function spaces (discrete or continuous), and thus linear combina-
tions of them can be used to approximate arbitrarily close to any nonlinear element in
that space. Multiplicative weighting, by raising the priorities of the alternatives to the
power of the priorities of the criteria (which it determines through additive
weighting!) then multiplying the results, has four major flaws: (a) It does not give
back weights of existing same ratio scale measurements on several criteria as it
should; (b) It assumes that the matrix of judgments is always consistent, thus
sacrificing the idea of inconsistency and how to deal with it, and not allowing
redundancy of judgments to improve validity about the real world; (c) Most criti-
cally, it does not generalize to the case of interdependence and feedback, as the AHP
generalizes to the Analytic Network Process (ANP), so essential for the many
decision problems in which the criteria and alternatives depend on each other; (d) It
always preserves rank which leads to unreasonable outcomes and contradicts the
many counterexamples that show rank reversals should be allowed; (6) Rank
preservation and reversal can be shown to occur without adding or deleting criteria,
such as by simply introducing enough copies of an alternative or for numerous other
reasons; this leaves no doubt that rank reversal is as intrinsic to decision making as
rank preservation also is; it follows that any decision theory must have at least two
modes of synthesis; in the AHP they are called the distributive and ideal modes, with
guidelines for which mode to use; rank can always be preserved by using the ideal
mode in both absolute measurement and relative measurement; (7) Group judg-
ments must be integrated one at a time carefully and mathematically, taking into
consideration when desired the experience, knowledge, and power of each person
involved in the decision, without the need to force consensus, or to use majority or
other ordinal ways of voting; the theorem regarding the impossibility of constructing
a social utility function from individual utilities that satisfies four reasonable
conditions which found their validity with ordinal preferences is no longer true when
cardinal ratio scale preferences are used as in the AHP. Instead, one has the possi-
bility of constructing such a function. To deal with a large group requires the use of
questionnaires and statistical procedures for large samples.
26 2 The Seven Pillars of the Analytic Hierarchy Process
A ratio is the relative value or quotient a/b of two quantities a and b of the same kind;
it is called commensurate if it is a rational number, otherwise it is incommensurate.
A statement of the equality of two ratios a/b and c/d is called proportionality. A ratio
scale is a set of numbers that is invariant under a similarity transformation (multi-
plication by a positive constant). The constant cancels when the ratio of any two
numbers is formed. Either pounds or kilograms can be used to measure weight, but
the ratio of the weight of two objects is the same for both scales. An extension of this
idea is that the weights of an entire set of objects whether in pounds or in kilograms
can be standardized to read the same by normalizing. In general if the readings from
a ratio scale are aw*i , i = 1,…,n, the standard form is given by wi = aw*i /aw*i = w*i /
w*i as a result of which we have wi = 1, and the wi, i = 1,…,n, are said to be
normalized. We no longer need to specify whether weight for example is given in
pounds or in kilograms or in another kind of unit. The weights (2.21, 4.42) in pounds
and (1, 2) in kilograms, are both given by (1/3, 2/3) in the standard ratio scale form.
The relative ratio scale derived from a pairwise comparison reciprocal matrix of
judgments is derived by solving:
X
n
aij wj ¼ kmax wi ð2:1Þ
j¼1
X
n
wi ¼ 1 ð2:2Þ
i¼1
with aji = 1/aij or aij aji = 1 (the reciprocal property), a ij [ 0 (thus A is known as
a positive matrix) whose solution, known as the principal right eigenvector, is
normalized as in (2.2). A relative ratio scale does not need a unit of measurement.
When aij ajk = aik, the matrix A = (aij) is said to be consistent and its principal
eigenvalue is equal to n. Otherwise, it is simply reciprocal. The general eigenvalue
formulation given in (2.1) is obtained by perturbation of the following consistent
formulation:
A1 An
2 3 2 3 2 3
A1 w1 w1
w1 w1
6 w1 ... wn 7 6 7 6 7
6 7 6 7 6 7
. 6 .. 7 6 .. 7 6 . 7
Aw ¼ .. 6 ... ..
. .7 6 . 7 ¼ n6 .. 7 ¼ nw:
6 7 6 7 6 7
4 wn 5 4 5 4 5
w1 . . . wwnn
An wn wn
where A has been multiplied on the right by the transpose of the vector of weights
w = (w1,…,wn). The result of this multiplication is nw. Thus, to recover the scale
from the matrix of ratios, one must solve the problem Aw = nw or (A - nI)w = 0.
This is a system of homogeneous linear equations. It has a nontrivial solution if and
2.2 Ratio Scales 27
Zb
k Kðs; tÞwðtÞdt ¼ wðsÞ ð2:4Þ
a
Zb
wðsÞds ¼ 1 ð2:5Þ
a
where instead of the matrix A we have as a positive kernel, K(s,t) [ 0. Note that the
entries in a matrix depend on the two variables i and j which assume discrete values.
Thus the matrix itself depends on these discrete variables, and its generalization, the
kernel function also depends on two (continuous) variables. The reason for calling it
kernel is the role it plays in the integral, where without knowing it we cannot determine
the exact form of the solution. The standard way in which (2.3) is written is to move the
eigenvalue to the left hand side which gives it the reciprocal form. In general, by abuse
of notation, one continues to use the symbol k to represent the reciprocal value. Our
equation for response to a stimulus is now written in the standard form (2.4) with the
normalization condition (2.5). Here also, we have the reciprocal property (2.6) and as
in the finite case, the kernel K(s,t) is consistent if it satisfies the relation (2.7):
K ðs; tÞK ðt; sÞ ¼ 1 ð2:6Þ
In the discrete case, the normalized eigenvector was independent of whether all
the elements of the pairwise comparison matrix A are multiplied by the same
constant a or not, and thus we can replace A by aA and obtain the same eigen-
vector. Generalizing this result we have:
K ðas; atÞ ¼ aK ðs; tÞ ¼ kðasÞ=kðatÞ ¼ a kðsÞ=kðtÞ
which means that K is a homogeneous function of order one. In general, when
f (ax1,…,axn) = an f (x1,…,xn) holds, f is said to be homogeneous of order n. Because
K is a degenerate kernel, we can replace k(s) above by k(as) and obtain w(as). We
2.2 Ratio Scales 29
have now derived from considerations of ratio scales the following condition to be
satisfied by a ratio scale:
Theorem A necessary and sufficient condition for w(s) to be an eigenfunction
solution of Fredholm’s equation of the second kind, with a consistent kernel that is
homogeneous of order one, is that it satisfy the functional equation
wðasÞ ¼ bwðsÞ
where b = aa.
We have for the general damped periodic response function w(s),
log b log s log s
wðsÞ ¼ Ce P
log a log a
Instead of assigning two numbers wi and wj and forming the ratio wi/wj we assign
a single number drawn from the fundamental 1–9 scale of absolute numbers to
represent the ratio (wi/wj)/1. It is a nearest integer approximation to the ratio wi/wj.
The derived scale will reveal what the wi and wj are. This is a central fact about the
relative measurement approach of the AHP and the need for a fundamental scale.
In 1846 Weber found, for example, that people while holding in their hand dif-
ferent weights, could distinguish between a weight of 20 g and a weight of 21 g, but
could not if the second weight is only 20.5 g. On the other hand, while they could not
distinguish between 40 and 41 g, they could between 40 and 42 g, and so on at higher
levels. We need to increase a stimulus s by a minimum amount Ds to reach a point
where our senses can first discriminate between s and s ? Ds. Ds is called the just
noticeable difference (jnd). The ratio r = Ds/s does not depend on s. Weber’s law
states that change in sensation is noticed when the stimulus is increased by a constant
percentage of the stimulus itself. This law holds in ranges where Ds is small when
compared with s, and hence in practice it fails to hold when s is either too small or too
large. Aggregating or decomposing stimuli as needed into clusters or hierarchy levels
is an effective way for extending the uses of this law.
In 1860 Fechner considered a sequence of just noticeable increasing stimuli. He
denotes the first one by s0. The next just noticeable stimulus is given by
30 2 The Seven Pillars of the Analytic Hierarchy Process
D s0
s1 ¼ s1 þ Ds0 ¼ s0 þ s ¼ s0 ð1 þ rÞ
s0 0
based on Weber‘s law.
Similarly s2 ¼ s1 þD s1 ¼ s1 ð1 þ rÞ ¼ s0 ð1 þ r Þ2 s0 a2 : In general
sn ¼ sn1 a ¼ s0 an ðn ¼ 0; 1; 2; . . .Þ:
Thus stimuli of noticeable differences follow sequentially in a geometric pro-
gression. Fechner noted that the corresponding sensations should follow each other
in an arithmetic sequence at the discrete points at which just noticeable differences
occur. But the latter are obtained when we solve for n. We have n ¼ ðlog nloga
s log s0 Þ
and
sensation is a linear function of the logarithm of the stimulus. Thus if M denotes the
sensation and s the stimulus, the psychophysical law of Weber–Fechner is given by
M ¼ a log s þ b; a 6¼ 0:
We assume that the stimuli arise in making pairwise comparisons of relatively
comparable activities. We are interested in responses whose numerical values are
in the form of ratios. Thus b = 0, from which we must have log s0 = 0 or s0 = 1,
which is possible by calibrating a unit stimulus. Here the unit stimulus is s0. The
next noticeable stimulus is s1 ¼ s0 a ¼ a which yields the second noticeable
response a log a. The third noticeable stimulus is s2 ¼ s0 a2 which yields a
response of 2a log a. Thus we have for the different responses:
M0 ¼ a log s0 ; M1 ¼ a log a; M2 ¼ 2a log a; . . .; Mn ¼ na log a:
While the noticeable ratio stimulus increases geometrically, the response to that
stimulus increases arithmetically. Note that M0 = 0 and there is no response. By
dividing each Mi by M1 we obtain the sequence of absolute numbers 1, 2, 3,… of the
fundamental 1–9 scale. Paired comparisons are made by identifying the less domi-
nant of two elements and using it as the unit of measurement. One then determines,
using the scale 1–9 or its verbal equivalent, how many times more the dominant
member of the pair is than this unit. In making paired comparisons, we use the nearest
integer approximation from the scale, relying on the insensitivity of the eigenvector
to small perturbations (discussed below). The reciprocal value is then automatically
used for the comparison of the less dominant element with the more dominant one.
Despite the foregoing derivation of the scale in the form of integers, someone might
think that other scale values would be better, for example using 1.3 in the place of 2.
Imagine comparing the magnitude of two people with respect to the magnitude of one
person and using 1.3 for how many there are instead of 2.
We note that there may be elements that are closer than 2 on the 1–9 scale, and we
need a variant of the foregoing. Among the elements that are close, we select the
smallest. Observe the incremental increases between that smallest one and the rest of
the elements in the close group. We now consider these increments to be new
elements and pairwise compare them on the scale 1–9. If two of the increments are
themselves closer than 2 we treat them as identical, assigning a 1 (we could carry this
on ad infinitum—but we will not). In the end each component of the eigenvector of
2.3 Paired Comparisons and the Fundamental Scale 31
In Fig. 2.1, an unripe cherry tomato is eventually and indirectly compared with a
large watermelon by first comparing it with a small tomato and a lime, the lime is
then used again in a second cluster with a grapefruit and a honey dew where we then
divide by the weight of the lime and then multiply by its weight in the first cluster,
and then use the honey dew again in a third cluster and so on. In the end we have a
32 2 The Seven Pillars of the Analytic Hierarchy Process
comparison of the unripe cherry tomato with the large watermelon and would
accordingly extended the scale from 1–9 to 1–721.
Such clustering is essential, and must be done separately for each criterion. We
should note that in most decision problems, there may be one or two levels of clusters
and conceivably it may go up to three or four adjacent ranges of homogeneous
elements (Maslow put them in seven groupings). Very roughly we have in decreasing
order of importance: (1) Survival, health, family, friends and basic religious beliefs
some people were known to die for; (2) Career, education, productivity and lifestyle;
(3) Political and social beliefs and contributions; (4) Beliefs, ideas, and things that are
flexible and it does not matter exactly how one advocates or uses them. Nevertheless
one needs them, such as learning to eat with a fork or a chopstick or with the fingers as
many people do interchangeably. These categories can be generalized to a group, a
corporation, or a government. For very important decisions, two categories may need
to be considered. Note that the priorities in two adjacent categories would be suffi-
ciently different, one being an order of magnitude smaller than the other, that in the
synthesis, the priorities of the elements in the smaller set have little effect on the
decision. We do not have space to show how some undesirable elements can be
compared among themselves and gradually extended to compare them with desir-
able ones as above. Thus one can go from negatives to positives but keep the mea-
surement of the two types positive, by eventually clustering them separately.
2.5 Clustering and Using Pivots to Extend the Scale from 1–9 to 1–? 33
Let H be a complete hierarchy with h levels. Let Bk be the priority matrix of the
kth level, k = 2,…, h. If W0 is the global priority vector of the pth level with
respect to some element z in the (p - 1)st level, then the priority vector W of the
qth level (p \ q) with respect to z is given by the multilinear (and thus non-
linear) form,
W ¼ Bq Bq1 . . . Bpþ1 W 0 :
The global priority vector of the lowest level with respect to the goal is given by,
W ¼ Bh Bh1 . . . B2 W 0 :
Fig. 2.4 Combining the two costs through additive or multiplicative syntheses
If desired, one can include a remainder term to estimate the error. With regard to
additive and multiplicative syntheses being close, one may think that in the end it
does not matter which one is used, but it does. Saaty and Hu [7] have shown that
despite such closeness on every matrix of consistent judgments in a decision, the
synthesized outcomes by the two methods not only lead to different final priorities
(which can cause a faulty allocation of resources) but more significantly to
different rankings of the alternatives. For all these problems, but more significantly
because it does not generalize to dependence and feedback even with consistency
guaranteed, and because of the additive nature of matrix multiplication needed to
compute feedback in network circuits to extend the AHP to the ANP, I do not
recommend ever using multiplicative synthesis. It can lead to an undesirable
ranking of the alternatives of a decision.
Given the assumption that the alternatives of a decision are completely indepen-
dent of one another, can and should the introduction (deletion) of new (old)
alternatives change the rank of some alternatives without introducing new
(deleting old) criteria, so that a less preferred alternative becomes most preferred?
Incidentally, how one prioritizes the criteria and subcriteria is even more important
than how one does the alternatives which are themselves composites of criteria.
Can rank reverse among the criteria themselves if new criteria are introduced?
Why should that not be as critical a concern? The answer is simple. In its original
form utility theory assumed that criteria could not be weighted and the only
important elements in a decision were the alternatives and their utilities under the
various criteria. Today utility theorists imitate the AHP by rating, and some even
by comparing the criteria, somehow. There was no concern then about what would
happen to the ranks of the alternatives should the criteria weights themselves
36 2 The Seven Pillars of the Analytic Hierarchy Process
change as there were none. The tendency, even today, is to be unconcerned about
the theory of rank preservation and reversal among the criteria themselves.
The house example of the previous section teaches us an important lesson. If
we add a fourth house to the collection, the priority weights of the criteria Price
and Remodeling Cost would change accordingly. Thus the measurements of the
alternatives and their number, which we call structural factors, always affect the
importance of the criteria. When the criteria are incommensurate and their
functional priorities are determined in terms of yet higher level criteria or goals,
one must still weight such functional importance of the criteria by the structural
effect of the alternatives. What is significant in all this is that the importance of the
criteria always depends on the measurements of the alternatives. If we assume that
the alternatives are measured on a different scale for each criterion, it becomes
obvious that normalization is the instrument that provides the structural effect to
update the importance of the criteria in terms of what alternatives there are.
Finally, the priorities of the alternatives are weighted by the priorities of the
criteria that depend on the measurements of the alternatives. This implies that the
overall ranking of any alternative depends on the measurement and number of all
the alternatives. To always preserve rank means that the priorities of the criteria
should not depend on the measurements of the alternatives but should only derive
from their own functional importance with respect to higher goals. This implies
that the alternatives should not depend on the measurements of other alternatives.
Thus one way to always preserve rank is to rate the alternatives one at a time. In
the AHP this is done through absolute measurement with respect to a complete set
of intensity ranges with the largest value intensity value equal to one. It is also
possible to preserve rank in relative measurement by using an ideal alternative
with full value of one for each criterion.
The logic about what can or should happen to rank when the alternatives
depend on each other has always been that anything can happen. Thus, when the
criteria functionally depend on the alternatives, which implies that the alternatives,
which of course depend on the criteria, would then depend on the alternatives
themselves, rank may be allowed to reverse. The Analytic Network Process (ANP)
is the generalization of the AHP to deal with ranking alternatives when there is
functional dependence and feedback of any kind. Even here, one can have a
decision problem with dependence among the criteria, but with no dependence of
criteria on alternatives and rank may still need to be preserved. The ANP takes
care of functional dependence, but if the criteria do not depend on the alternatives,
the latter are kept out of the supermatrix and ranked precisely as they are dealt with
in a hierarchy [8].
Examples of rank reversal abound in practice, and they do not occur because
new criteria are introduced. The requirement that rank always be preserved or that
it should be preserved with respect to irrelevant alternatives. To every rule or
generalization that one may wish to set down about rank, it is possible to find a
counterexample that violates that rule. Here is the last and most extreme form of
four variants of an attempt to qualify what should happen to rank given by Luce
and Raiffa, each of which is followed by a counterexample. They state it but and
2.7 Rank Preservation and Reversal 37
then reject it. The addition of new acts to a decision problem under uncertainty
never changes old, originally non-optimal acts into optimal ones. The all-or-none
feature of the last form may seem a bit too stringent… a severe criticism is that it
yields unreasonable results. The AHP has a theory and implementation procedures
and guidelines for when to preserve rank and when to allow it to reverse. One
mode of the AHP allows an irrelevant alternative to cause reversal among the
ranks of the original alternatives.
The distributive mode of the AHP produces preference scores by normalizing the
performance scores; it takes the performance score received by each alternative
and divides it by the sum of performance scores of all alternatives under that
criterion. This means that with the Distributive mode the preference for any given
alternative would go up if we reduce the performance score of another alternative
or remove some alternatives.
The Ideal mode compares each performance score to a fixed benchmark such as
the performance of the best alternative under that criterion. This means that with the
Ideal mode the preference for any given alternative is independent of the perfor-
mance of other alternatives, except for the alternative selected as a benchmark. Saaty
and Vargas [11] have shown by using simulation, that there are only minor differ-
ences produced by the two synthesis modes. This means that the decision should
select one or the other if the results diverge beyond a given set of acceptable data.
The following guidelines were developed by Millet and Saaty [3] to reflect the
core differences in translating performance measures to preference measures of
alternatives. The Distributive (dominance) synthesis mode should be used when the
decision maker is concerned with the extent to which each alternative dominates all
other alternatives under the criterion. The Ideal (performance) synthesis mode
should be used when the decision maker is concerned with how well each alternative
performs relative to a fixed benchmark. In order for dominance to be an issue the
decision-maker should regard inferior alternatives as relevant even after the ranking
process is completed. This suggests a simple test for the use of the Distributive
mode: if the decision maker indicates that the preference for a top ranked alternative
under a given criterion would improve if the performance of any lower ranked
alternative was adjusted downward, then one should use the Distributive synthesis
mode. To make this test more actionable we can ask the decision maker to imagine
the amount of money he or she would be willing to pay for the top ranked alternative.
If the decision maker would be willing to pay more for a top ranked alternative after
learning that the performance of one of the lower-ranked alternatives was adjusted
downward, then the Distributive mode should be used.
Consider selecting a car: Two different decision makers may approach the same
problem from two different points of views even if the criteria and standards are the
same. The one who is interested in ‘‘getting a well performing car’’ should use the Ideal
38 2 The Seven Pillars of the Analytic Hierarchy Process
mode. The one who is interested in ‘‘getting a car that stands out’’ among the alter-
natives purchased by co-workers or neighbors, should use the Distributive mode.
Here we consider two issues in group decision making. The first is how to
aggregate individual judgments, and the second is how to construct a group choice
from individual choices.
Let the function f(x1, x2,…, xn) for synthesizing the judgments given by n judges,
satisfy the
1. Separability condition (S): f(x1, x2,…,xn) = g(x1)g(x2)… g(xn) for all x1,
x2,…,xn in an interval P of positive numbers, where g is a function mapping P
onto a proper interval J and is a continuous, associative and cancellative
operation. [(S) means that the influences of the individual judgments can be
separated as above.]
2. Unanimity condition (U): f(x, x,…,x) = x for all x in P. [(U) means that if all
individuals give the same judgment x, that judgment should also be the syn-
thesized judgment.]
3. Homogeneity condition (H): f(ux1, ux2,…,uxn) = uf(x1, x2,…,xn) where u [ 0
and xk, uxk (k = 1,2,…,n) are all in P. [For ratio judgments (H) means that if all
individuals judge a ratio u times as large as another ratio, then the synthesized
judgment should also be u times as large.]
4. Power conditions (Pp): f(xp1,xp2,…,xpn) = fp(x1, x2,…,xn). [(P2) for example
means that if the kth individual judges the length of a side of a square to be xk,
the synthesized judgment on the area of that square will be given by the square
of the synthesized judgment on the length of its side.]
Special case (R = P-1): f(1/x1, 1/x2,…,1/xn) = 1/f(x1, x2,…,xn). [(R) is of par-
ticular importance in ratio judgments. It means that the synthesized value of the
reciprocal of the individual judgments should be the reciprocal of the synthesized
value of the original judgments.]
Aczel and Saaty (see [9, 10]) proved the following theorem:
Theorem The general separable (S) synthesizing functions satisfying the una-
nimity (U) and homogeneity (H) conditions are the geometric mean and the root-
mean-power. If moreover the reciprocal property (R) is assumed even for a single
n-tuple (x1, x2,…,xn) of the judgments of n individuals, where not all xk are equal,
then only the geometric mean satisfies all the above conditions.
2.8 Group Decision Making 39
In any rational consensus, those who know more should, accordingly, influence
the consensus more strongly than those who are less knowledgeable. Some people
are clearly wiser and more sensible in such matters than others, others may be
more powerful and their opinions should be given appropriately greater weight.
For such unequal importance of voters not all g’s in (S) are the same function. In
place of (S), the weighted separability property (WS) is now: f(x1,
x2,…,xn) = g1(x1)g2(x2)… gn(xn). [(WS) implies that not all judging individuals
have the same weight when the judgments are synthesized and the different
influences are reflected in the different functions (g1, g2,…,gn).]
In this situation, Aczel and Alsina (see [9]) proved the following theorem:
Theorem The general weighted-separable (WS) synthesizing functions with the
unanimity (U) and homogeneity (H) properties are the weighted geometric mean
q q
f ðx1 ; x2 ; . . .; xn Þ ¼ x11 x22 . . . xqnn and the weighted root-mean-powers
p ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
f ðx1 ; x2 ; . . .; xn Þ ¼ c q1 xc1 þ q2 xc2 . . . þ qn xcn ; where q1 ? q2 ? … ? qn = 1,
qk [ 0 (k = 1,2,…,n), c [ 0, but otherwise q1,q2,…,qn, c are arbitrary constants.
If f also has the reciprocal property (R) and for a single set of entries
(x1,x2,…,xn) of judgments of n individuals, where not all xk are equal, then only
the weighted geometric mean applies. We give the following theorem which is an
explicit statement of the synthesis problem that follows from the previous results,
and applies to the second and third cases of the deterministic approach:
ðiÞ ðiÞ
Theorem If x1 ; . . .; xn i = 1,…, m are rankings of n alternatives by m inde-
pendent judges and if ai is the importance of judge i developed
m fromamhierarchy
Pm Q ai Q ai
for evaluating the judges, and hence ai ¼ 1; then xl ; . . .; xn are
i¼1 i¼1 i¼1
the combined ranks of the alternatives for the m judges.
The power or priority of judge i is simply a replication of the judgment of that
judge (as if there are as many other judges as indicated by his/her power ai), which
implies multiplying his/her ratio by itself ai times, and the result follows.
The first requires knowledge of the functions which the particular alternative
performs and how well it compares with a standard or benchmark. The second
requires comparison with the other alternatives to determine its importance.
Given a group of individuals, a set of alternatives (with cardinality greater than (2),
and individual ordinal preferences for the alternatives, Arrow proved with his
Impossibility Theorem that it is impossible to derive a rational group choice
(construct a social choice function that aggregates individual preferences) from
ordinal preferences of the individuals that satisfy the following four conditions,
i.e., at least one of them is violated:
40 2 The Seven Pillars of the Analytic Hierarchy Process
Bibliography
3.1 Introduction
This chapter illustrates the use of the Analytic Hierarchy Process in determining
the amount and location of space assigned to each room, according to its function,
in the design of a house [1].
We develop the plan of a house to satisfy a family’s needs by considering the
size of the lot, the size and shape of the different architectural spaces, their pri-
orities, and their overall contiguity. We begin by identifying needs and then
integrating them with the final plan by relating shape, size, geometric design and
location to our mental criteria and personal needs. This approach also permits one
to treat all these needs and their relation to the environment in a coherent
framework.
We study design by addressing six basic factors: (1) architectural needs; (2)
Budget allocation; (3) allocation of areas to satisfy needs; (4) size and shape of the
areas; (5) clustering spaces according to general needs; (6) identifying and locating
individual spaces in each cluster.
We assume that in this hypothetical example we are dealing with the needs of
a family of three, the husband (H), the wife (W), and their child (CH). They
already own a fully paid lot, on which they wish to build a custom designed
house (see Fig. 3.1). The husband and wife have a maximum total disposable
budget of $70,000 to cover construction and landscaping improvement costs.
Other costs such as legal fees, architectural fees, permits, etc. are not included.
The cost of construction is assumed to be $30 per square foot and that of
landscaping $1.20 per square foot. (The matrices of judgments are included for
completeness.)
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 41
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_3, Ó Springer Science+Business Media New York 2012
42 3 Architectural Design
The members of the family are considered to be the actors or decision makers. The
parents will exercise far greater influence than the child on the design process.
Since each of them will be seeking to satisfy similar human needs expressed
3.2 Architectural Needs 43
through architectural design, we will consider their objectives to be the same, the
difference being that each party may assign a different weight to each objective.
These objectives are: (1) The need to eat (M); (2) The need to rest (R); (3) The
need to entertain (E); (4) The need to clean (K); (5) The need to store (S); (6)The
need to communicate, e.g., through halls and entrance, within the house (C).
First we develop a set of priorities indicating the relative power of the
parties in determining the outcome for the above architectural needs (see
Table 3.1). Table 3.1 consists of calculations done for the parties. The table
shows that the child will not influence decisions greatly, even if the husband
were to enter a full coalition with him. In our estimate in this hypothetical
example, combining both parties would only increase their joint power slightly
(0.453), but they would still be unable to surpass the power of the wife (0.547).
Because of this fact, we eliminate the child from further consideration of
influence, and focus on the husband (H) and wife (W) as the sole represen-
tatives of the family needs.
Next, we prioritize the importance of the objectives from the standpoints of W
and H (Table 3.2). We then weight each outcome by the power of the corre-
sponding party (see the last column of Table 3.2), and take the sum of the com-
posite outcomes (see Table 3.2a and Fig. 3.2). This sum will be used to allocate
square footage to each architectural space for each objective.
Next, a predetermined array of architectural spaces are assigned to each party,
arranged in such a way as to satisfy one or more of the above objectives. For
example, the family room can be placed in such a way as to satisfy entertainment
and resting needs (see Fig. 3.2). The parties will be allowed to change or modify
their estimate of the importance of these functions. We then pairwise compare
these spaces with respect to the six objectives listed above, to determine what each
party’s preference is to satisfy its particular objectives. This yields weights or
priorities for the space assigned. This prioritization can be revised after later
negotiations among the parties, in order to reach an acceptable compromise
solution. The results are shown in Tables 3.3 and 3.4. We note from these com-
parisons, that H and W have widely differing preferences for the space assigned
within each objective. This is reflected in their priorities for the spaces also shown
in the tables. These preferences are then weighted by the priority of the objective
for that party and by the power of the party, i.e., we use the last column of
Table 3.2 to obtain the last columns of Tables 3.3 and 3.4. Table 3.5 gives the sum
of the priorities of the spaces for H and W.
44 3 Architectural Design
From the total budget of $70,000, H and W must decide on the proportion they
would like to allocate to construction and to exterior improvement. With the
assistance of an architect, they both agree to allocate 85% or nearly $60,000 to
Construction and the remainder to Landscaping.
We now proceed to allocating a proportion of the area to each objective. With
the $60,000 budget and construction cost of $30 per sq. ft., we calculate that the
house would have a maximum area of 2000 sq. ft. Table 3.2a gives the parties’
joint preference priorities of the importance of each objective. These priorities are
now used to proportionately allocate the 2000 sq. ft. among the objectives, as
shown in Table 3.6 below.
Next, we use each party’s preferences to assess the size of the architectural space
associated with each objective (determined previously in Table 3.5). Here the
question we answer through pairwise comparisons is as follows: How does a party
3.3 Allocation of the Budget, Property and Architectural Spaces 45
evaluate the size of a particular space, as compared with the size of other spaces
within the same objective, in order to obtain a fair split of spaces for that objective?
The preferences are shown in Tables 3.7 and 3.8. There, we also weight the resulting
priorities by the power of each party (see last columns of Tables 3.7 and 3.8).
The combined overall weight for each space in each objective is the sum of
these for each party, as shown in Table 3.9. The last column of Table 3.9 gives the
space allocated to each architectural space within each objective, by multiplying
the previous column entries by the total square footage allotted to each objective
from Table 3.6.
3.4 Dimensions
The choice of dimensions for the rooms (measured in feet) can be made by
hierarchical analysis if we have no idea of what is wanted. However, in general
people formulate notions about these needs from the environment in which they
live. They would make adjustments on what they already know and have expe-
rienced. To unfold that complexity of feelings in a hierarchy would be superfluous.
Thus, the shape and dimensions of the rooms given in this chapter is an expression
of our feelings as what a family of three would require and how each choice would
fit with the dimensions of other rooms, whose dimensions are known, to obtain
efficient use of each space and proper balance in functionality and aesthetics. For
46 3 Architectural Design
example, a very large living room and very small bedrooms or dining room would
probably be considered to be misproportioned.
As an illustration we take the ‘‘Rest’’ objective which is allotted 433 sq. ft.
Three types of spaces or rooms are associated with it. They and their comparisons
according to relative contribution to ‘‘Rest’’ are shown in Tables 3.7, 3.8 and 3.9.
The priorities and resulting proportion of the total area are given by:
Architectural space Objective’s area Weight Area allocated to space (Sq. Ft.)
MB Master bedroom 433 0.664 288
B-2 Bedroom 2 433 0.139 60.3
SR Study room 433 0.197 85.3
Our goal now is to let both H and W determine the shape of different ‘‘Rest’’
areas, check for optimum size, and revise inconsistencies such as the one above,
where B-2 should presumably have a higher weight than the SR. For each
objective, we proceed according to the priority of the spaces within it by selecting
the higher one, then the next higher and so on.
The first pairwise comparison was done on the Master Bedroom (see
Table 3.10), given that this space is the most important with respect to the ‘‘Rest’’
objective. The dimensions developed through prioritization were 140 by 200
yielding 280 sq. ft. (less than the 288 sq. ft. allotted to this space). We must show
how this is done. In the following pairwise comparisons, we assume that the
parties had previously determined a set of dimensions to be ranked. To determine
this set, H and W could use, as a reference point, their experience with currently
known and used spaces, whose dimensions could accurately be measured. They
could then make judgments accordingly. For example, suppose that H and W
presently use a Master Bedroom whose dimensions are 140 by 140 . Moreover, let us
suppose that they feel comfortable with a width of 140 but not with a 140 length.
They would then vary the length to make the comparisons as shown in Table 3.10.
It is clear that their preferred dimensions may not fit harmoniously with the total
area for each objective. In this case adjustments must be made to produce a good
48 3 Architectural Design
fit and by reducing or increasing the area allotted to that space. They would then
examine whether this allocation is compatible with the priorities of the next space
within the objective and the availability of remaining space to allocate to it. For
example, MB is 140 9 200 and B-2 is 140 9 140 , exceeding the area allotted to R
(i.e., 433 sq. ft.). Thus SR could not be included. Later we also make adjustments
in the areas allotted to the objectives, to remedy such violations. For example, this
can be achieved in a second iteration, where the parties H and W change their
preferences to a MB of 160 9 160 which yields a total area of 256 sq. ft., and a B-2
of 140 9 140 which when combined with the MB yield a total area of 424 sq. ft.
assigned to the R objectives. This leaves 9 sq. ft. (433 sq. ft.—424 sq. ft.) as a
residual area in the R objective.
This process can be continued for each of the remaining objectives, until a
final list containing the dimensions of all spaces is achieved, as shown in
Table 3.11. The last column of this table gives the surplus (or deficit) so far
produced for each objective. The total area must be no more than 2000 sq. ft.
This is attained by adjusting the dimensions of some spaces and by revising the
priorities when it is clear that a serious judgment error has been committed and
that the comparison made is unacceptable to each of the parties. People usually
start out with high aspirations and end up making compromises.
Some of the inconsistencies are a result of disagreements between the parties
as to what should be valued most. The comparisons are sometimes unrealistic
and are adjusted in terms of what is needed apart from high ideals for variety. In
the final analysis, the dimensions of each of the spaces are slightly adjusted to
absorb a deficit and to yield a surplus to meet the overall constraint of space
availability.
To determine the location and contiguity of the architectural spaces, we first set
down criteria for location. We note that the front view of a house usually faces
the street and may be separated from it by a garden. The other three sides
separate the house from neighbors with spaces of varying size, which may be
used for different purposes, e.g., driveway, garden or the like. For the house
under consideration, the dimensions are shown in Fig. 3.1, where the sur-
rounding space is an open grassy area with a tree on the right hand side of the
house. Note that the construction site must fall within the 4000 square foot
rectangle in Fig. 3.1. Our task is to indicate the location of the objectives and
then the spaces within this area.
52 3 Architectural Design
We do not insist that there should be straight rectangular walls surrounding the
construction site. Thus, the contiguity of the rooms and hence the general layout of
the house is our next task, and is the most difficult. We focus on the three highest
priority objectives: E,R and M. They have 84.2% of the weight. We divide the
priority of each by their total, so that the three sum to unity.
Next we identify the location criteria used to determine the relative importance
of the different locations in this single level house (see Fig. 3.3). They are: Front
View (FV) (facing yard and street); Back View (BV) (facing yard); Left Side View
(LV); and Right Side View (RV) (with a tree).
We then develop priorities for the relative importance of these locations with
respect to the objectives E, R and M. The parties may perform the following types
of pairwise comparisons shown in Table 3.12 below, which answer the following
question: Which location is most desirable for E or R or M?
Graphically, these priorities are illustrated in Fig. 3.4. Now we turn our
attention to locating the highest priority room, the Living Room, in the highest
priority location for Entertainment needs, which is in the Front View. To
accomplish this task, the building area is divided into quadrants of 40 9 40 as
shown in Fig. 3.5, and the Husband and Wife may jointly (or separately if desired)
attempt to position LR. This prioritization process shows which quadrants are
chosen for the Living Room. Table 3.13 shows these calculations.
As we observe from this pairwise comparison, quadrant C6–10H6–10 was
selected for the LR. Therefore, we proceed to lay it out in the lot as shown in
Fig. 3.6.
Next, the parties position the second highest priority space in the house, the FR.
Because this space satisfies E needs, it would also have to be positioned towards
the front part of the lot as previously determined. Table 3.14 shows the calcula-
tions for the FR.
The third highest priority space is the LB, which is also the last space we must
place within the E objective. Following the same procedure explained above, the
parties determine its position by pairwise comparisons, as shown in Table 3.15.
The fourth ranking space was the DR, which now satisfies M needs. The parties
now use the right hand side of the construction site to position this particular space.
The same procedure will be applied to the next space in ranking, which happens to
be the Master Bedroom, but this space will be placed towards the back part of the
lot since it satisfies R needs. The positioning of this space is shown in Table 3.16
below and in Fig. 3.6.
56 3 Architectural Design
Alternative positions for bathrooms and closets, for example, can be similarly
analyzed and located with respect to their bedrooms.
The readjustments involve iterations of the process to obtain a coherent
architectural design. If the problem of locating rooms in a single level dwelling
3.5 Contiguity of Architectural Spaces 59
seems complicated, more than that of a two or more story house, it is even more
complicated for a split-level house. The main reason for this is that one must not
only locate rooms in the two levels, but must also match the levels at the
boundary where they meet to obtain aesthetic, efficient and practical transition
from one to the other. Split-level houses are known to provide the opportunity
for greater efficiency and variety in using space that a two story house. Inclined
ground is better suited to split-level housing than flat ground.
3.6 Conclusion
We have illustrated the use of a new method that can be used as an aid to
formalize some of the intuitive aspects of architectural design. Through its use,
the architect and the builder can enable owner participation (at the desired level)
in the process.
Bibliography 61
Bibliography
1. Saaty TL, Beltran MH (1982) The analytic hierarchy process: a new approach to deal with
fuzziness in architecture. Architectural Sci Rev 25:64–69
Chapter 4
Why is the Principal Eigenvector
Necessary?
4.1 Introduction
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 63
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_4, Ó Springer Science+Business Media New York 2012
64 4 Why is the Principal Eigenvector Necessary?
the real world. When we deal with tangibles, a pairwise comparison judgment
matrix may be perfectly consistent but irrelevant and far off the mark of the true
values. For several reasons a modicum of inconsistency may be considered as a
good thing and forced consistency without knowledge of the precise values as an
undesirable compulsion. If one insists on consistency, people would be required to
be like robots, unable to change their minds with new evidence, and unable to look
within for judgments that represent their thoughts, feelings and preferences.
The AHP also uses a principle of hierarchic composition to derive composite
priorities of alternatives with respect to multiple criteria from their priorities with
respect to each criterion. It consists of multiplying each priority of an alternative by
the priority of its corresponding criterion and adding over all the criteria to obtain the
overall priority of that alternative. This is perhaps the simplest way for composing
priorities. The additive approach is also crucial in doing composition using the
limiting powers of a priority rather than a judgment matrix when dependence and
feedback are considered in decision-making. Different methods for deriving priori-
ties within the same hierarchy can lead to different final values for the alternatives [6].
Now we ask the question, what is priority or more generally what meaning should we
attach to a priority vector of a set of alternatives? We can think of two meanings. The
first is a numerical ranking of the alternatives that indicates an order of preference
among them. The other is that the ordering should also reflect intensity or cardinal
preference as indicated by the ratios of the numerical values and is thus unique to
within a positive multiplicative constant (a similarity transformation). It is the latter
that interests us here as it relates to the principle of hierarchic composition under a
single criterion. Given the priorities of the alternatives and given the matrix of
preferences for each alternative over every other alternative, what meaning do we
attach to the vector obtained by weighting the preferences by the corresponding
priorities of the alternatives and adding? It is another priority vector for the alter-
natives. We can use it again to derive another priority vector ad infinitum. Even then
what is the limit priority and what is the real priority vector to be associated with the
alternatives? It all comes down to this: What condition must a priority vector satisfy
to remain invariant under the hierarchic composition principle? A priority vector
must reproduce itself on a ratio scale because it is ratios that preserve the strength of
preferences. Thus a necessary condition that the priority vector should satisfy is not
only that it should belong to a ratio scale, which means that it should remain invariant
under multiplication by a positive constant c, but also that it should be invariant under
hierarchic composition for its own judgment matrix so that one does not keep getting
new priority vectors from that matrix. In sum, a priority vector x must satisfy the
relation Ax = cx, c [ 0. We will show that as a result of the need for invariance to
produce a unique priority vector, x must be the principal right eigenvector of A and
c is its corresponding principal eigenvalue. Our problem for positive reciprocal
matrices and their priorities is a special case of the following:
4.2 What is a Priority Vector? 65
Theorem For a given positive matrix A, the only positive vector x and only
positive constant c that satisfy Ax = cx, is a vector x that is a positive multiple of
the Perron vector (principal eigenvector) of A, and the only such c is the Perron
value (principal eigenvalue) of A.
Proof We know that the (right) Perron vector and Perron value satisfy our
requirements. We also know that the algebraic multiplicity of the Perron value is
one, and that there is a positive left eigenvector of A (call it z) corresponding to the
Perron value. Suppose there is a positive vector y and a (necessarily positive)
scalar d such that Ay = dy. If d and c are not equal, then by biorthogonality [2] y is
orthogonal to z, which is impossible since both vectors are positive. If d and c are
equal, then y and x are dependent since c has algebraic multiplicity one, and y is a
positive multiple of x.
It is also true that if one starts with any priority vector and transforms it through
multiplication by A any number of times, in the limit, the product converges to the
Perron vector of A. Significantly and interestingly, for our purpose to derive pri-
orities for a special type of positive matrices, the foregoing theorem can also be
shown to hold for a class of positive reciprocal matrices that are consistent and
near consistent without the use of the theorem of Perron. We know that the
principal eigenvector is the priority vector of a consistent matrix. For such a matrix
aij ¼ wi =wj ; and it follows from Aw = nw that the vector w ¼ ðw1 ; . . .; wn Þ that is
also the principal eigenvector of A is its priority vector with corresponding prin-
cipal eigenvalue c = n. We can show by small and continuous perturbation [3, 8]
of a consistent matrix A that the resulting near consistent matrix (see the next
section), has its priority vector as its principal eigenvector obtained as a pertur-
bation of the corresponding principal eigenvector of A. Thus if we assume that a
judgment matrix is obtained as a small perturbation of an underlying consistent
matrix constructed from a ratio scale w ¼ ðw1 ; . . .; wn Þ; its priority vector coin-
cides with its principal eigenvector obtained as a small perturbation of w. For the
perturbation proof, which is fairly long and elaborate, see [5].
That would end our quest if we could also say what to do about a positive
inconsistent matrix with large inconsistency. We need to improve its consistency
to speak of its priority vector. Using the Perron vector and Perron root of such a
matrix, we show that it can be transformed in steps to a near consistent matrix
whose priority vector we now know is its principal eigenvector.
of the form aij þ cij ; a reciprocal perturbation aij eij ; eji ¼ e1
ij is multiplicative. It
can be transformed to an additive perturbation of a consistent matrix by writing:
wi wi wj wj 1
þ cij ¼ eij ; eij ¼ 1 þ c ; eji ¼ e1
ij ¼ þ cji ¼ w :
wj wj wi ij wi 1 þ wji cij
Note that with a reciprocal perturbation we ensure that kmax n which helps
determine the validity of w as a priority vector of a near consistent matrix. We
have
X
n X
eij ¼ j
aij wj =wi ¼ ½Awi =wi ¼ kmax wi =wi ¼ kmax :
j¼1
The computation
P
n Pn P
n P
n
nkmax ¼ ð eij Þ ¼ eii þ ðeij þ eji Þ
i¼1 j¼1 i¼1 i;j¼1
i6¼j
X
n
¼nþ ðeij þ e1 2
ij Þ n þ ðn nÞ=2 ¼ n
2
i;j¼1
i6¼j
reveals that kmax n: Moreover, since x þ 1=x 2 for all x [ 0, with equality if
and only if x = 1, we see that kmax ¼ ngij if and only if all cij = 1, which is
equivalent to having all aij = wi/wj. The foregoing arguments show that a positive
reciprocal matrix A has kmax n; with equality if and only if A is consistent.
To improve the validity of the priority vector, we must transform a given reci-
procal judgment matrix to a near consistent matrix. In practice, the judgments
available to make the comparisons may not be sufficient to bring the matrix to near
consistency. In that case we abandon making a decision based on the information
we have, and must seek additional knowledge to modify the judgments.
The final question then is how to construct the c perturbations in a general
reciprocal matrix. A judgment matrix already has some built in consistency; it is
not an arbitrary reciprocal matrix. Among others, inconsistency in a matrix may be
due to an error such as putting aji instead of aij in the i,j position which if
appropriately detected and changed the matrix may become near consistent or at
least improve the consistency of A. Because the principal eigenvector is necessary
for representing dominance (and priorities when near consistency is obtained), we
must use an algorithm based on the eigenvector, whose existence is assured by
4.4 The General Case: How to Transform a Positive Reciprocal Matrix 67
Table 4.1 A family’s house buying pairwise comparison matrix for the criteria
Size Trans Nbrhd Age Yard Modern Cond Finance w v
Size 1 5 3 7 6 6 1/3 1/4 0.173 0.047
Trans 1/5 1 1/3 5 3 3 1/5 1/7 0.054 0.117
Nbrhd 1/3 3 1 6 3 4 6 1/5 0.188 0.052
Age 1/7 1/5 1/6 1 1/3 1/4 1/7 1/8 0.018 0.349
Yard 1/6 1/3 1/3 3 1 1/2 1/5 1/6 0.031 0.190
Modern 1/6 1/3 1/4 4 2 1 1/5 1/6 0.036 0.166
Cond 3 5 1/6 7 5 5 1 1/2 0.167 0.059
Finance 4 7 5 8 6 6 2 1 0.333 0.020
8max = 9.669
Consistency ratio (CR) = 0.17
We conclude that
@kmax
¼ vi wj a2ji vj wi for all i; j ¼ 1; . . .; n:
@aij
Because we are operating within the set of positive reciprocal matrices,
@kmax
¼ @k@amax
@aji ij
for all i and j. Thus, to identify an entry of A whose adjustment
within the class of reciprocal matrices would result in the largest rate of change in
kmax we should examine the n(n - 1)/2 values fvi wj a2ji vj wi g; i [ j and select
(any) one of largest absolute value. This is the method proposed for positive
reciprocal matrices by Harker [1].
To illustrate the methods discussed above, consider an example involving the
prioritization of criteria used to buy a house for a family whose members coop-
erated to provide the judgments (Table 4.1).
Table 4.2 gives the array of partial derivatives for the matrix of criteria in
Table 4.1.
The (4, 8) entry in Table 4.2 (in bold print) is largest in absolute value. Thus, the
family could be asked to reconsider their judgment (4, 8) of Age versus Finance.
68 4 Why is the Principal Eigenvector Necessary?
One needs to know how much to change a judgment to improve consistency, and we
show that next. One can then repeat this process with the goal of bringing the C.R.
within the desired range. If the indicated judgments cannot be changed fully
according to one’s understanding, they can be changed partially. Failing the
attainment of a consistency level with justifiable judgments, one needs to learn
more before proceeding with the decision.
Two other methods, presented here in order of increasing observed efficiency in
practice, are conceptually different. They are based on the fact that
X
n
nkmax n ¼ ðeij þ e1
ij Þ:
i;j¼1
i6¼j
This suggests that we examine the judgment for which cij is farthest from one,
that is, an entry aij for which aij wj/wi is the largest, and see if this entry can
reasonably be made smaller. We hope that such a change of aij also results in a
new comparison matrix with that has a smaller Perron eigenvalue. To demonstrate
how improving judgments works, take the house example matrix in Table 4.1. To
identify an entry ripe for consideration, construct the matrix cij = aij wj/wi
(Table 4.3). The largest value in Table 4.3 is 5.32156, which focuses attention on
a37 = 6.
How does one determine the most consistent entry for the (3, 7) position?
Harker has shown that when we compute the new eigenvector w after changing the
(3, 7) entry, we want the new (3, 7) entry to be w3/w7 and the new (7, 3) to be w7/w3.
On replacing a37 by w3/w7 and a73 by w7/w3 and multiplying by the vector w one
obtains the same product as one would by replacing a37 and a73 by zeros and the two
corresponding diagonal entries by two (see Table 4.4).
4.4 The General Case: How to Transform a Positive Reciprocal Matrix 69
We take the Perron vector of the latter matrix to be our w and use the now-
known values of w3/w7 and w7/w3 to replace a37 and a73 in the original matrix.
The family is now invited to change their judgment towards this new value of a37
as much as they can. Here the value was a37 = 0.102/0.223 = 1/2.18, approxi-
mated by 1/2 from the AHP integer valued scale and we hypothetically changed it
to 1/2 to illustrate the procedure (see Table 4.5). If the family does not wish to
change the original value of a 37, one considers the second most inconsistent
judgment and repeats the process.
A refinement of this approach is due to W. Adams. One by one, each reciprocal
pair aij and aji in the matrix is replaced by zero and the corresponding diagonal
entries aii and ajj are replaced by 2, the principal eigenvalue kmax is then computed.
The entry with the largest resulting kmax is identified for change as described
above. This method is in use in the Analytic Network Process (ANP) software
program Superdecisions [4].
4.5 Conclusions
the inconsistency is less than or equal to a desired value [7]. We also mentioned
three ways and illustrated two of them, as to how to improve the consistency of
judgments and transform an inconsistent matrix to a near consistent one.
Bibliography
1. Harker PT (1987) Derivatives of the perron root of a positive reciprocal matrix: with
applications to the analytic hierarchy process. Appl Math Comput 22:217–232
2. Horn RA, Johnson CR (1985) Matrix analysis. Cambridge University Press, New York
3. Lancaster P, Tismenetsky M (1985) The theory of matrices, 2nd edn. Academic, New York
4. Saaty TL (2001) Decision making with dependence and feedback: the analytic network
process. RWS Publications, 4922 Ellsworth Ave. Pittsburgh, PA, p 15213
5. Saaty TL (2001) Decision making with the AHP: Why is the principal eigenvector necessary?’’
Proceedings of the sixth international symposium on the analytic hierarchy process, Berne-
Switzerland, 2–4 August 2001
6. Saaty TL, Hu G (1998) Ranking by the eigenvector versus other methods in the analytic
hierarchy process. Appl Math Lett 11(4):121–125
7. Saaty TL, Vargas L (1984) Inconsistency and rank preservation. J Math Psychol 28(2):205–
214
8. Vargas LG (1983) Analysis of sensitivity of reciprocal matrices. Applied mathematics and
computation, 12, Elsevier Science Publishing Co Inc, New York, pp 301–320
Chapter 5
Designing a Mousetrap
5.1 Introduction
This chapter illustrates the sequential use of hierarchies in the selection, design and
marketing of a mouse trap. The project described in this chapter is not so much
about building a better mousetrap as it is about a person’s hope to reach a higher
state by engulfing oneself in the project. Not a higher state in the classical sense
that is indicative of a great mind expanding, but a state that helps a person to grow
through doing, always reminding one that there are numerous ways to see a
problem. By reminding the student to be open and sensitive to new ideas and new
creative processes, the professor can foster positive growth. Emphasizing that you
never reach a state where you are all that there is, and your ways are the best, can
lead to a healthy respect for continued growth buttressed by a positive attitude, and
a striving for a more fulfilling life.
A better mousetrap was not the goal of the assignment. To develop a person
through the use of reasoning and agonizing can lead to a sense of accomplishment
that comes only when the inputs come from within the person. The project has
raised my self-esteem, while helping to humble me so that I can use my unique
gifts while respecting the gifts of others in the healthiest of manners. A very fine
line exists between self-confidence and arrogance when a person is elevated in this
way, however, the sense of humility that comes from respecting the ideas of others
can hold a check over the entire process.
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 71
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_5, Ó Springer Science+Business Media New York 2012
72 5 Designing a Mousetrap
Table 5.2 Judgements and priorities with respect to goal for optimal shock delivery
Head Trunk Neck Leg-feet Tail
Head 3.0 2.0 1/4 5.0
Trunk 1/2 1/5 4.0
Neck 1/4 4.0
Leg-feet 5.0
Tail
0.216
Head XXXXXXXXXXXXXXX
0.099
Trunk XXXXXXX
0.142
Neck XXXXXXXXXX
0.497
Leg-feet XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
0.047
Tail XXXXXX
shock to the mouse’s legs and feet (and even any other body parts). As can be seen
from the first three models, the decisions made in each model drive those made in
the subsequent model. With that in mind it was clear how to kill the rodent, what
area should receive the shock and the optimal means to do it.
The next issue was how to attract the mouse to the trap. For this a fourth model
was developed whose goal was to identify the best luring device (Fig. 5.3).
Criteria included cost, with a set of sub-criteria; ease of use, also with sub-criteria;
appeal (attracting power), with its own sub-criteria; and reusability. The alternative
means of attracting a mouse were developed and evaluated under each set of
criteria and sub-criteria.
The synthesis concluded that the alternatives of food scent and sex scent
were within 0.001 of each other. For this reason it was decided to include both
lures in the mousetrap. Combining food scent and sex scent would be much
more powerful than attracting the rodent by means of any type of real food or
dim lighting system (Table 5.4).
5.4 The Trap Shape 75
The fifth model was used to decide on the optimal shape of the trap given that the
trap was to perform electrically (Fig. 5.4).
The alternatives generated included a sphere, a tube, a platform, a box and a
dome. Each alternative was evaluated for its design appeal based on the presumed
human need for an attractive trap. From the rodent’s perspective, the trap needed
natural appeal which focused on the mouse’s curious nature. Other criteria were
materials needed in construction and the various considerations about ease of
manufacture (Table 5.5).
Once the synthesis was completed, with an inconsistency ratio of 0.05, the obvious
clear shape of choice was a tube design with a composite priority of 0.479. Up to
this point it has been clear how each previous model drove the subsequent model’s
creation. Now, with all the elements in place, an idea born of the AHP could begin
76 5 Designing a Mousetrap
to take form. Before entering the actual design phase, however, it was important to
evaluate the costs and benefits of actual mousetraps. Two models were developed;
one for the costs (Fig. 5.5) of the trap and the other for the benefits (Fig. 5.6). The
mousetrap costs model looked at both using a trap and not using it. Under not
using a trap further considerations included health issues, property damage, lower
property value and the potential embarrassment of having rodents in the house.
Sub-criteria under using a trap were the purchasing cost, the operating cost, trap
maintenance time, the visibility of the trap in the house. Under the sub-criteria of
not using the trap (i.e., health issue and property damage) were food damage,
mouse’s excrement in the house, the bite factor and damage considerations to
clothes and furniture.
5.5 The Cost and Benefits of the Trap 77
With the synthesis completed it became clear that the costs of having the trap
outweighed those of not having the trap. Attention now focused on the benefits of
using the trap given the findings of the last model (Table 5.6).
The benefits model, similar to the costs model, looked at both using and not
using the mousetrap. Under using the trap, the issues included a clean home
environment, maintaining current property value, no damage to food/clothes or
property, a rodent free environment and no embarrassment due to the lack of a
mouse. Considered under not using the trap were the opportunity to earn money by
breeding and selling them, development of mouse races and the chance to obtain
low cost pets. The mousetrap benefits model showed that the benefits of using a
trap outweighed those of not using one and a subjective judgment followed that the
benefits did outweigh the costs (Table 5.7).
At this point it seemed that only the trap design itself was left to be developed, but
a question arose as to whether this trap, once developed, could prosper monetarily.
Our concern now focussed on profits and on the development of a model (Figs. 5.7
and 5.8) that would assist in deciding which markets should be entered once the
trap took physical form. Criteria included the trap’s profit potential, customer need
and geographic location. Both profit potential and customer need used sub-criteria
78 5 Designing a Mousetrap
of high, medium and low with greatest weight going to high. Geographic location
had sub-criteria of sections of the country (i.e., Eastern, Midwest, Central,
Mountain and Western). Under these were the alternatives of metro cities, metro
suburbs and rural areas. Once the synthesis was completed the first choice of a
market to enter was found to be metro suburbs at 0.549, followed by metro cities at
0.266 and rural at 0.185 (Tables 5.8 and 5.9).
5.7 The Design 79
With regard to the actual trap, some of the design questions included the shape of
the entrance to the trap, how long should the trap be, where to place wire griding,
how to make it reusable and a number of other questions. Figure 5.9 contains the
preliminary drawings of the front and back of the trap and one complete sideview.
Also, a picture of the front and back with the wire grid floor is included to aid in
understanding the design.
Figures 5.10, 5.11, 5.12 and 5.13 contain the formal design plans. Figure 5.10
gives an overview of the trap. The 3 1/400 section outlines the inner chamber while
the 4 1/400 section shows the outlays of the chamber. All solid lines depict the outer
layer of steel and the inner dashed line depicts the thickness of the materials used.
This convention is adhered to for all the following exhibits.
Figure 5.11 contains views of the front and rear ends. If the trap was ‘‘elec-
trified’’ at all times, it was unlikely that a mouse would enter. A means of detecting
entry was needed to send a signal to the delay mechanism housing assembly.
80 5 Designing a Mousetrap
front back
At 1.2 s after the beam is broken, the four corners of the wire griding receive an
electrical charge of 110 V. The trap chamber is the section that the mouse enters
while the electrical components enclosure opens to dispose of the dead rodent with
a twist of the wrist. This is why a seven inch piano hinge, and the magnetic closure
are included. Remember the first hierarchy.
Reusability and ease of use were indeed among the initial criteria. The lure was
located on the top section of the trap chamber to make replacement easy and to force
the mouse’s weight downward on the wire grid as it pushed up to get toward the lure.
Figure 5.12 is the rear elevation, the view of the trap from its magnetic closure
side.
Figure 5.13 contains two views, a front elevation and Section A–A.
The front elevation is the hinge side. The overall length of the complete trap is
900 long while the section that opens is 700 long. While observing the front
5.7 The Design 81
elevation, note the two downward pointing arrows labeled ‘‘A’’. This represents a
cross-section of this area and a top down view can be seen in section A–A. This is
the complete 900 electrical wire grid that is located in the channel of the trap
chamber.
It is important to note that this type of thinking and design can come only when
you ask not only how it has been done, but how you might do it differently.
82 5 Designing a Mousetrap
5.8 Conclusions
The AHP was used to model the stages of designing a new mouse trap. The model
helps elucidate the thinking process. Without the beginning models as stepping
stones, the later models would be meaningless separate entities with no tie to the
overall goal of building a better mousetrap.
Table 5.10 contains an overall view of the reconstituted master hierarchy. Note
that this exhibit displays only the first three levels of the total picture. To
understand the complete model, it is useful to examine the previous tables and
visualize where each component part fits.
This exercise (the overall model) demonstrates how decisions in one model lead
to the creation of another model and affect choices made for each model thereafter.
The cost/benefit analysis and the decision about what markets to enter were driven
by the preceding models but not to the same extent that the first five models were.
Although the master hierarchy model was created after the completion of the
project, it serves to demonstrate the sequence of events followed and the relative
importance of each with respect to those models they affected.
Bibliography
1. Dobias AP (1990) Designing a mousetrap using the analytic hierarchy process and expert
choice. Eur J Oper Res 48(1):57–65
2. Saaty TL (1988) Decision making for leaders. RWS Publications, Pittsburgh
Chapter 6
Designing the Best Catamaran
6.1 Introduction
This chapter illustrates use of the Analytic Hierarchy Process in the selection and
design of a sailboat. How does one bring together one’s ideas when designing a
versatile sailing machine? Imagination plays an important role but is full of
disconnected thoughts. The AHP was used to first choose the overall sailboat
design, and then to select some fundamental hydrodynamic features of the newly
designed boat.
There are four basic sailboat designs: sloop, ketch, catamaran and trimaran. The
sloop has one hull and one mast. The ketch has one hull with one large mast
located at the centerpoint of the yacht’s deck, and a smaller mast (mizzenmast)
that is usually located behind the helm (where the captain steers). Each vessel uses
one mainsail and one jib (triangular shaped foresail), except for the ketch which
also uses a mizzensail (mainsail on the mizzenmast).
A catamaran and a trimaran differ in only one major feature. The catamaran is
made of two pontoons that are attached by a variety of spar configurations. Usually
a tarp stretches between the two pontoons. The trimaran has three pontoons, with
the center one being the main body of the vessel. These pontoons are usually
connected by a reinforced fiberglass structure. Both the catamaran and the trimaran
have one mast, that is located around the centerpoint.
By thoroughly evaluating each type of vessel, the AHP led to the development
of a design comprising the best features of both the sloop and the catamaran. This
new vessel design is called a Main-Hull-Mono-maran (MHM-maran). The MHM-
maran broadly resembles a catamaran. It has two pontoons, but one is much larger
than the other. The larger pontoon (major pontoon) is similar to a sloop in both size
and appearance. The smaller pontoon (minor pontoon), is approximately 1/3 the
length and 1/4 the width of the major pontoon. The minor pontoon is for added
sailing performance while the major pontoon provides all living and storage
facilities.
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 83
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_6, Ó Springer Science+Business Media New York 2012
84 6 Designing the Best Catamaran
Four criteria were used (see Fig. 6.1) to evaluate the four alternative designs:
catamaran, ketch, sloop and trimaran. With respect to sailing performance, both
the sloop and the ketch have a greater ability to sail close to the wind (close-hauled
position). But the marans are far lighter than their counterparts, thus they sail faster
in most ‘‘points’’ of sailing. By points of sailing we mean the direction from which
the wind is blowing: a run is with the wind blowing from behind, a reach is with
the wind blowing from either side and close-to-the-wind (close-hauled) is with the
wind blowing 30–35° off the bow.
The criteria were separated into two distinct categories: cruising and racing.
With respect to personal design preferences, long-distance racing was the criterion
that carried the greatest weight. Following it was short-distance racing, long-
distance cruising and short-distance cruising, in that order. The goal was to design
some sort of racing yacht.
Criteria Weights
1. Short dist. racing 0.231
2. Long dist. racing 0.623
3. Short dist. cruising 0.052
4. Long dist. cruising 0.094
Until recently, the marans did not receive much respect when it came to
structural integrity. They were not considered very seaworthy, and in storms the
pontoons tended to break apart. But with the development of expensive and strong
plastics, vast improvements have been made. Still, many sailors prefer the tradi-
tional structural strength of both the sloop and the ketch.
Evaluating each alternative with the above criteria revealed some interesting
results. The catamaran, as almost every sailor knows, is superior to all other yachts
when it comes to short-distance racing. It is very light and fast but it does not have
adequate storage or living facilities, so it is not suitable for long-distance racing.
Instead, the sloop was determined to be the best yacht for long-distance racing. It
really did not matter which yacht was best for short or long-distance cruising,
because neither of these criteria had significant weight to greatly affect the overall
outcome. The overall weights of the alternatives are given by:
6.2 Basic Design 85
Alternatives Weights
1. Sloop 0.379
2. Catamaran 0.348
3. Ketch 0.138
4. Trimaran 0.134
Notice how close the weights of both the sloop and the catamaran are. Up to
this point, the results from each hierarchy were used without question, mainly
because the final weights clearly marked which was the superior alternative.
Before beginning this evaluation it was assumed that the conclusion would be the
sloop. But this did not turn out to be the case. It was the catamaran. The catamaran
ended up following the sloop by only 0.031. This led to an entirely new conclu-
sion—the development of a sailboat that takes advantage of the distinct charac-
teristics of both the sloop and the catamaran.
Since a sloop consists of one large hull, and a catamaran consists of two
pontoons, a yacht that combines both characteristics will have to be somewhere
in between. Two options are possible—a yacht with one main hull and a single
fixed pontoon or a yacht with a variable pontoon. This is where the concept of
the Main Hull Mono-maran enters. It is a distinct yacht that has one large hull
(main hull) that is shaped much like a sloop’s hull. It also has an attached
pontoon which provides the vessel with characteristics typically associated with a
catamaran. When the sea is rough, one has the option to have the pontoon trail
the vessel.
There are only two alternatives in this hierarchy, Yacht A and Yacht B. The
only difference between them is that A has a fixed pontoon whereas B has a
variable one which can move from the port to the starboard side and vice versa.
There are three criteria: Agility (0.558), Stability (0.122) and Anchorability
(0.320).
AGILITY is the ability of a design to quickly maneuver under sail. STABILITY
is the ability to keep the yacht heeled as upright as possible. Finally, ANCHOR-
ABILITY refers to how easy it is to, or not to, anchor or dock the yacht. A fixed
pontoon limits docking ability because the dock (slip) needs to be sufficiently wide
to accommodate the vessel. The final weights for the two designs are given by:
Alternatives Weights
Yacht B 0.696
Yacht A 0.304
86 6 Designing the Best Catamaran
Thus, in this hierarchy, Yacht B, the model with variable pontoon, is decidedly
preferred.
Now that a basic yacht concept has been derived through the process provided
by the AHP, the specifics of the MHM-maran, namely, the keel, the rudder, and the
sails were designed.
The keel is the most important hydrodynamic feature of any sailboat. A keel is
responsible for a boat’s stability and prevents it from sliding across the water on
certain points of sailing (close-hauled & reach position). Once a keel is chosen, the
rudder is relatively easy to design. As the hydrodynamics of the rudder are depen-
dent upon those of the keel: there is a close relationship between keel and rudder.
Since the Australians introduced a winged keel in the 1983 America’s Cup,
there has been a tremendous amount of research in this area. Many of the keel test
results are not available to the public, since it is mainly the very competitive
yachting syndicates that perform and sponsor these activities.
When selecting a keel, performance and practicality were considered
(see Fig. 6.2). The criteria were: mud, rock, sand and the degree of heel of the yacht.
The mud, rock and sand elements were selected, because these are the substances
that sailors most often hit when they go aground. A keel provides stability and many
other performance factors, but in this case only the degree of heel (leaning to one
side) was taken into consideration. The two criteria (performance and practicality)
were considered equally important, and within practicality, the three substances the
keel can come in contact with are also considered equally relevant.
Thus, we have:
Criteria Weights
Heel 1/2
Mud 1/6
Rock 1/6
Sand 1/6
Keels are classified into three basic types—winged, torpedo and fin. A deep
keel was not considered because such keels have been outdated and are poor
performers. Generally speaking, a winged keel provides the least amount of heel
because of the lift action that each wing creates as it glides through the water.
The wings actually serve as a vertical stabilizer. As the yacht heels (leans on its
side), the wings keep the boat a few more degrees upright. The benefit of such a
feature is that it allows the vessel to catch as much wind as possible, while
minimizing the amount spilled. However, a winged keel is unsuitable for the
MHM-maran, because the minor pontoon will serve the same purpose. When the
minor pontoon is correctly positioned on the leeward side of the vessel, it will
6.4 The Keel 87
effectively reduce the vessel’s degree of heel. When evaluating the criteria, the
wing was the least favored.
The torpedo keel outperforms the fin keel with the respect to heeling because
the torpedo limits the amount of heel. The torpedo keel is given that name
because it is a standard fin keel with a torpedo shaped bulb on the bottom
(Fig. 6.3a). It is no longer a common type of keel, but at one time it was thought
that having much of the lead weight at the base of the keel would act as a
stabilizer. Some designers have combined the features of both the winged and
torpedo keels.
The most common keel is the fin keel (Fig. 6.3b). It goes straight down deep
and tends to draw much water. Variations of the fin keel exist, such as a shoal draft
keel which draws less water, but is thicker and longer. The shoal draft keel was not
considered because it is not a good hydrodynamic performer. The fin keel was
favored and after placing it on the major pontoon, it was found to be an excellent
choice. When sailing in a close-hauled or reach position, the fin keel would reduce
the MHM-marans sliding effect across the water.
The winged keel presents a problem in mud (Fig. 6.3c). If one runs into mud,
there is a good chance of getting stuck. The wings tend to anchor themselves into
the mud, making it difficult to break away. The torpedo keel is no worse than the
fin keel, but the bulb base can also stick into the mud.
Rocks are a problem for all keels. Sailing in shallow waters with a rocky
bottom is hazardous. A winged keel does not fare well when it hits rock—
structurally it can crack, or a wing can break off if the impact is sufficiently
strong. A torpedo keel is better in rock than a winged keel but no better than a
fin keel. A fin keel usually receives the least damage when its base hits rock. If
the forward middle part of a keel is hit by rock, it could suffer chips and cracks.
When a boat hits sand, it goes ‘‘thump, thump’’ a sensation sometimes asso-
ciated with engine problems. Fin and torpedo keels do about the same thing when
hitting sand, causing the boat to skid across. With a winged keel, one needs to be
concerned as to whether or not enough thump is due to one of the wings, which
could break off. We have:
88 6 Designing the Best Catamaran
Alternatives Weights
Fin 0.391
Torpedo 0.352
Wing 0.257
We determined that the fin keel is the best suited for our purposes. Earlier, we
spoke of its hydrodynamic advantages. With respect to the environment, the fin
keel is the least affected by different bottom surfaces. It is relatively easy to get
out of the mud, and it is least damaged when coming into contact with hard
surfaces.
There is a definite relationship between a keel and a rudder and once the keel type
is selected, the rudder more or less falls into place. What had to be considered was
the rudder work coefficient (CWR) and the angle of deflection. These became the
criteria. There are three basic alternatives—an attached rudder, a skeg type and a
spade rudder (see Fig. 6.4).
The attached rudder is connected to the keel with hinges (Fig. 6.5a). At one
time, this type of rudder was widely used, especially when the yacht had a deep
keel (a keel that began shortly after the bow and went all the way back to the aft).
Neither deep keels nor attached rudders are in much use nowadays.
The skeg rudder is made in a variety of shapes and sizes (Fig. 6.5b). It is not
attached to the keel. Rather, it is attached to a skeg that extends from the base of the
hull. The theory behind such a design is that it is more streamlined and allows water to
pass slightly faster by the rudder, thus making the rudder’s movements more effective.
The spade rudder is the sailing industry’s most commonly used rudder
(Fig. 6.5c). It is attached to a rudder post which in turn is connected to the helm
(steering wheel) by a pulley assembly mechanism. It is also made in a variety of
shapes and sizes. Some go deep, but never deeper than the deepest point of the
keel, while others tend to be less rectangular. This type tends to be the number one
choice among yacht manufacturers today, mostly because it is easy to configure
hydrodynamically with specific keel characteristics. Thus, it was not surprising
that the spade rudder was the most favored. Spade rudders are commonly used in
conjunction with fin keels.
Fig. 6.5 Types of rudders a Deep keel with an attached rudder. b Skeg type rudder. c Spade
rudder
So far we have selected the major underwater hydrodynamic features and the basic
overall design (see Fig. 6.6). Now it is time to take a closer look. The big design
problem is connecting the pontoons together. There are only two choices at this
stage: one solid piece, or two poles (see Fig. 6.7).
Each choice had to be flexible enough, so that it would move with the pontoon
to the other side of the vessel. A solid connection, that leaves no visible gap
between both pontoons, raises the question of seaworthiness. Such a structure
could get swamped by a wave and break. Also, a solid structure would weigh more
than two poles, thus adding to the overall weight of the vessel.
By comparing both alternatives with respect to the criteria: the poles were most
favored except on strength. Using poles is not only less expensive, but easier to
design. On the main pontoon, a dual track assembly would follow the rear
perimeter on which the ends of the poles would be attached. On the minor pon-
toon, two curved tracks would travel along the centerline where the poles would
also connect (Fig. 6.8).
This design connects the two poles, and the tracks serve to provide the poles
with movement, allowing the minor pontoon to swing around the main hull
(Figs. 6.9 and 6.10).
The poles are strong, present no problem with respect to seaworthiness, and
would be much lighter than a solid structure. Poles can also extend and retract,
thus they are most suitable for the necessary pontoon swings.
Next we selected the material from which the pontoons would be built
(see Fig. 6.11). The possibilities are fiberglass, aluminum, titanium and stainless
steel. These materials happened to be the most resistant to rust. Titanium is light,
strong but also very expensive and difficult to work with. Yet, it is a popular metal
among the yacht racing community, because of its lightness and strength.
Aluminum alloys are much cheaper than titanium and are sufficiently strong for
most yachting needs. Aluminum is light, but much heavier than titanium. Fiber-
glass is good for salt water, but it is also heavy. It cannot undergo much direct
stress. Stainless steel is relatively expensive and heavy, yet it is very strong.
6.6 The Overall Mhm-Maran Structure 91
By having a dual track assembly follow the aft half of the major pontoon’s toe
rail (see Fig. 6.12) the minor pontoon would be mobile. It would be able to swing
around to the leeward side (the side where the wind exits the surface of the boat) to
minimize the MHM-maran degree of heel. On the run position, it would not matter
which side the minor pontoon is positioned, since the wind is blowing from
92 6 Designing the Best Catamaran
behind. The evaluation of all materials with respect to strength, cost and weight
gave rise to titanium as the most favored. This guarantees that a strong, and
flexible connection would be made between both pontoons.
6.7 Conclusion
Bibliography
7.1 Introduction
This chapter illustrates the use of the AHP for selecting the most appropriate
bridge design in two different applications. The one recommended in the second
application coincides with the decision that was actually made, demonstrating that
the exclusion of an important, but hard to perceive actor, can alter the final
decision. In this example we learn that decision making is not simply including
multiple criteria in the decision, but more importantly the diverse people or groups
who influence the outcome because of their own purposes.
By the close of the 1990s, the Golden Triangle Area in the City of Pittsburgh
will once again be under construction to improve the flow of traffic into, out of,
and through the city. The Mon Warf will be reconstructed, and the Fort Pitt Bridge
will undergo rehabilitation. The Commonwealth Bridge Project is an effort by
federal, state and local agencies to construct an alternative route across the
Monongahela River to alleviate traffic congestion between the central business
district of downtown Pittsburgh and Pittsburgh International Airport (Fig. 7.1).
A new bridge, and a high occupancy vehicle (HOV) busway will be built to the
Wabash Tunnel [3]. The bridge will consist of three HOV lanes and a lane for
pedestrian traffic. The Port Authority and PennDOT want to begin construction by
1997, so that the new bridge will be open for use when the Fort Pitt Bridge is
closed for rehabilitation. The cost of the entire busway project is estimated at
$300 million, and estimates for the new bridge range from $30 to $40 million.
At the time of the writing of the two reports1 on which this chapter is based, the
bridge type had not been decided, although the second paper (a follow up on the
first) did point to the bridge type that was chosen a few weeks later. The difference
between the two approaches is the shift in emphasis among the stakeholders from
1
See acknowledgments in the Preface.
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 95
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_7, Ó Springer Science+Business Media New York 2012
96 7 The Selection of a Bridge
the Public to the Coast Guard because of concern with safety and with the
economic impact that the bridge would have on the area.
The bridge types considered by The Port Authority of Allegheny County (PAT)
were: a Cable-stayed bridge, a Truss bridge, and a Tied-arch bridge.
Although the original concept of cable-stays dates back to Egyptian sailing ships,
on which inclined ropes hanging from a mast were used to support a beam [1],
it was not until the nineteenth century that the first cable-stayed bridge, the
Roeblings bridge, at Niagara Falls New York, was constructed [4]. Built in 1855, it
spanned 807 ft across a river gorge (Fig. 7.2).
Despite the fulfillment of this engineering feat, modern cable-stayed bridges were
not considered feasible until shortly after World War II, when German engineers
pioneered their development. The main hurdles were technology-based; the limita-
tion of high strength materials, structural analysis and construction methods [2].
7.2 Three Alternative Bridge Types 97
(a)
(b)
Fig. 7.3 a Cable-stay bridge with several cables. b Cable-stay bridge with cable separation
reduced
To overcome design limitations, only cables with very high tensile strength are
used. This minimizes beam deflection which becomes increasingly important as
the span increases. Moreover, adding several stay cables allows the use of more
slender deck beams (Fig. 7.3a) which require less flexural stiffness. By decreasing
the cable spacing supports (Fig. 7.3b), local bending moments in the girders are
also reduced.
The second presupposition for the success of cable-stayed bridge types is the
simplification of the deck cross-section. Simple double-edge girders supporting
transverse floor beams and top slabs provide a synergistic reinforcing action. As a
result, the deck structure acts as a unit over intermediate supports.
Because of their aesthetically pleasing form, cable-stayed bridges can be found
in almost all modern cities. The early cable-stayed bridges in North America were
mostly constructed from steel because that was the traditional method at the time.
98 7 The Selection of a Bridge
However, due to the high cost of material and labor the all-steel bridge has been
losing competitiveness. To address this decline, modern designers have developed
an orthotopic composite deck, a concrete deck slab supported by steel framing.
Today, virtually all cable-stayed bridges contain an orthotopic deck.
The economic viability and aesthetic appeal of the cable-stayed bridge make it
the most popular bridge type for spans ranging between 650 and 1650 ft. In North
America there were 25 cable-stayed bridges constructed or under development in
the 20 years between 1971 and 1993.
Truss bridges have been used in North America for decades. A truss may be
described as a triangulated assembly of straight members. The design of the truss
structure [2] allows applied loads to be resisted primarily by axial forces in its
straight truss members; the truss proper is loaded only at the nodes or intersection
of straight members. It is a very efficient and sturdy design.
Figure 7.4 shows five main types of trusses, each providing a slight variation in
load distribution. Also included in this figure is a general illustration of the Warren
Truss bridge. The two most common are the Pratt and the Warren designs. Bridge
design is typically performed on a case-by-case basis considering variables such as
traffic, durability, and dependability.
A truss bridge has two major structural advantages [2]: (a) The primary member
forces are axial loaded; (b) the open web system permits the use of a greater
overall depth than for an equivalent solid web girder. Both factors lead to economy
in material and a reduced dead weight. Increased depth results in reduced
deflections, and a more rigid structure. These advantages are achieved at increased
fabrication and maintenance costs.
The conventional truss bridge is typically most economical for medium spans
(500–1,500 ft). Traditionally it has been used for spans intermediate between the
plate girder and the stiffened suspension bridge designs. Modern construction
techniques have tended to increase the economical span of both steel and concrete
girders. Thus the steel truss bridge is a direct competitor to the cable-stayed for
intermediate spans. The relative lightness of a truss bridge is advantageous
because it can be assembled part-by-part using lifting equipment of small capacity.
Alternatively, the number of field connections may be supplanted by pre-assembly.
From an architectural perspective the truss bridge rarely possesses aesthetic
beauty. This is partly due to the complexity of the intersection of its load bearing
component. In bridges of moderate span, it appears best to provide a simple and
regular structure. For this reason, the Warren truss usually looks better than other
forms.
7.2 Three Alternative Bridge Types 99
(b)
(c)
(d)
(e)
The structural form of the arch has been used for its architectural beauty and
outstanding strength for centuries. Because of the manner in which it distributes
the applied load [2], with the aid of inward-acting horizontal components, the arch
is capable of distributing loads both above and below its structure. In a tied arch
design, the horizontal reactions to the arch rib are supplied by a tie at deck level.
Figure 7.5 illustrates common tied-arch bridge types.
Some of the distinguishable features of the tied-arch are that it reduces bending
moments in the superstructure and is fairly economical to construct relative to an
equivalent straight, simple supported girder or truss [2].
100 7 The Selection of a Bridge
(a)
(b)
(c)
Aesthetically, the arch has been the most appealing of all bridge types. Its
appearance is familiar and expressive. The curved shape is always pleasing to view.
The disadvantages of the tied-arch are probably its high relative fabrication and
building costs. The conventional curved arch rib usually entails the highest
expense. Building problems vary with structure type, with the least problematic
structure being the cantilever-arch and the tied-arch possibly being the most dif-
ficult. The difficulty with the latter arises from the fact that the horizontal reactions
are not available until the deck is completed.
The most desirable bridge type would conceivably be the one which brings the most
satisfaction to the greatest number of stakeholders. Using this goal, a hierarchy
(see Fig. 7.6) was developed with major stakeholders at the second level, the driving
criteria at the third level and the three alternative bridge types at the fourth level.
7.3.1 Stakeholders
Published reports have estimated the number of stakeholders involved in this project
to be in the hundreds. The most important stakeholders were aggregated into seven
broad groups. Commonality within the groups was maintained. The identifiable
groups are: a federal agency, the commercial business district, the public, state
7.3 The Decision Making Process 101
agencies, the Port Authority Transit, the designers, and special interest groups such
as concrete suppliers, steel manufacturers, environmentalists and others.
A Federal Agency (FHWA) represents an array of federal departments. They are
a key financier of the project and will have dictates with respect to the engineering
integrity of any bridge type.
The Commercial Business District (CBD) broadly represents the businesses in
downtown Pittsburgh. However, with respect to bridge type it is suggested that
Station Square is the dominant entity because of the interest in maintaining the
historical appearance of the site.
The Public represents the population of Pittsburgh which would use the new
service (and the bridge itself).
The Pennsylvania Department Of Transportation (PennDOT) represents the
complex interest of the state. These interests are financial (as the state provides
part of the capital), political, technical and environmental.
The Designers represent engineers, architects and planners and their repre-
sentative professional organizations. It is recognized that designers provide crucial
technical input and, as such, are strategically positioned to influence the decision
making process.
The Port Authority Transit (PAT) is the ultimate project owner. They are
responsible for all management issues from conception to construction, as well as
subsequent maintenance. This makes them a premier stakeholder.
Special Interest Groups (SIG) is a very broad category with diverse and pos-
sibly conflicting interests. With regard to the bridge type the three most significant
special interest groups are likely to be: concrete suppliers, steel manufacturers and
environmentalists. The indigenous steel industry of Pittsburgh has declined in size
and influence in recent times; however, the concrete industry remains strong.
Environmentalists are active and sometimes vocal.
102 7 The Selection of a Bridge
7.3.2 Criteria
In the level below the stakeholders are the criteria which drive the decision making
process. The most important criteria are:
Engineering Feasibility (EF): The technical knowledge and experience of both
the designers and contractors in regard to the bridge type.
Capital Cost (CC): Necessary funding.2 Because the costs were committed, low
costs are included in the overall benefits hierarchy as one the criteria.
Maintenance (MA): General cleaning, painting and inspection vary dramati-
cally with bridge type.
Aesthetics (AE): Architectural attractiveness.
Environmental Impact (EI): The ecological and historical adjustments that must
be compromised.
Durability (DU): The life of the bridge and the potential major repairs over and
above routine maintenance.
To find the most desirable bridge design, the actors were compared to determine
their relative influence (Table 7.1). The stakeholder ranking, as suggested by the
authors in order of importance, is PAT (0.337), CBD (0.221), Federal Agency
(0.136), State Agencies (0.136), Designers (0.085), Special Interests (0.056) and
the Public (0.029). The criteria were then compared according to each actor and
the composite relative priorities calculated (Table 7.2). Finally, the alternatives
were compared according to the criteria and the final priorities computed.
This information was synthesized to yield the most desirable bridge type which
was determined to be the Truss type bridge.
The priorities of the three alternatives are:
2
The cable-stayed bridge type is very efficient in terms of section sizes and material used.
However, there is limited experience with their use in the State of Pennsylvania. As a result, the
capital costswould vary greatly depending on the tender procedureused (statewide, nationwide or
international).
7.5 Bridge Selection Revisited 103
The analysis just presented was prepared by MBA students. Several months fol-
lowing the foregoing analysis another MBA student revisited the decision and sought
validation with two individuals closely involved in the bridge selection process. The
hierarchy developed in this second approach is given in Fig. 7.7 and the results are
provided in Table 7.3. Note that the hierarchy used is similar to the one given in
Fig. 7.6. However, the elements in the levels are not identical to the ones in the first
approach. The major difference between the two approaches is the addition of a new
stakeholder, the U.S. Coast Guard, and the deletion of The Public.
United States Coast Guard (USCG): River transportation has a significant
impact on the economy of Pittsburgh and the surrounding area in Western
Pennsylvania. Since there will be three bridges in close proximity to one another
(Smithfield Street, Monongahela River, and Fort Pitt), ample room must be
maintained for river traffic. The Coast Guard will also want to minimize the impact
of the bridge construction on river traffic.
Table 7.3 Priorities of the criteria for the revisited bridge decision
0.173 0.087 0.118 0.173 0.036 0.051 0.361
FHWA CBD USCG PennDOT Designers SIG PAT Priorities
EF 0.218 0.048 0.214 0.220 0.346 0.228 0.230 0.212
CC 0.354 0.137 0.043 0.344 0.248 0.100 0.121 0.195
MA 0.053 0.083 0.161 0.053 0.070 0.046 0.230 0.132
AE 0.073 0.400 0.067 0.114 0.155 0.432 0.121 0.146
EI 0.189 0.249 0.383 0.191 0.129 0.148 0.070 0.170
DU 0.112 0.083 0.123 0.078 0.052 0.046 0.230 0.142
7.6 Conclusion
The AHP was used to select the best alternative (the Tied-arch type bridge) from
among closely competing alternatives. It also facilitated the learning process and
gave users a more thorough understanding of the competing factors in a complex
decision making environment. It illustrates the sensitivity of the outcome to what
factors one chooses to include. It is certain that so far major public works divisions
7.6 Conclusion 105
have not been as thorough and comprehensive as may be needed to ensure the
success and longevity of an important project. More examples of such applications
need to be brought to the attention of authorities for better value and longer lasting
performance.
Bibliography
1. Leonhardt F, Zelbner W (1991) Past, present and future of cable-stayed bridges. In: Ito M et al.
(ed) Cable-stayed bridges, recent development and their future, Elsevier Science Publishers,
Amsterdam
2. O’Connor C (1971) Design of bridge superstructures. Wiley-Interscience, NY
3. Phase I: Airport busway/Wabash HOV (1992) Alternative analysis draft environmental impact
statement allegheny county, Pennsylvania. Prepared by U.S. Department of Transportation
Federal Transit Administration in cooperation with Federal Highway Administration, United
States Coast Guard, Pennsylvania Department of Transportation and Port Authority of
Allegheny County, Sept 1992
4. Tang MC (1991) Cable-stayed bridges in North America, In: Ito M et al. (ed) Cable-stayed
bridges, recent development and their future, Elsevier Science Publishers, Amsterdam
Chapter 8
Measuring Dependence Between
Activities: Input–Output Application
to the Sudan
8.1 Introduction
In this chapter we illustrate how to deal with dependence among the elements of
the same level of a hierarchy (inner dependence) with an application we made in
the design of the Transportation System for the Sudan (see [2]). The outcome of
this analysis was an input–output table which parallels the work of econometri-
cians. This application was done a few years after the Sudan Transport Study was
finished.
Input–output matrices in economics are generally obtained as follows. Let A1,
A2,…, AN denote N sectors of an economy and let S be a matrix whose sij entry
indicates the output from sector i which becomes an input to sector j. Let Yj be the net
PN
contribution from sector i to the final (consumer) demand. We have: sij ¼ Sj total
i¼1
intermediate output of sector j (domestic needs from other sectors) Sj þ Yj ¼ Oj total
s
output of sector j. The technological coefficients are obtained as follows Sj þij Yj ¼ wij
(contribution of sector i to produce a unit of output j)
sij sij Sj sij Sj
¼ ¼
S j þ Yj Sj S j þ Yj Sj Oj
To obtain the matrix of technological coefficients by the AHP we must estimate
sij/Sj and Sj/Oj. Let us see what these represent. Sj/(Sj ? Yj) represents the pro-
portion of the total output of sector j allocated to domestic consumption. The total
output is estimated, for j = 1,…, N, by means of the AHP by asking the following
question: How strong is one sector compared to another when allocating outputs to
domestic needs? If this question cannot be answered directly, domestic needs may
be hierarchically decomposed into production, demand, labor, capital, and cost.
These sectors are prioritized separately with respect to each criterion. After
prioritizing these criteria according to their impact on production, composition is
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 107
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_8, Springer Science+Business Media New York 2012
108 8 Measuring Dependence Between Activities: Input–Output Application
used to obtain an overall measure of importance for the sectors. Let us denote
the estimates of (Sj/Oj) by xj.
Again sij/Sj represents the proportion of the total intermediate output from
sector i allocated to sector j. We have
X
N
sij = Sj ¼ 1
i¼1
8.2 Application
then separately compare the aggregate with agriculture and transport and use the
resulting weight of the aggregate to compose the relevant weights from the four
sectors in the aggregate itself. To save space we have not written out justifications of
the judgments, which are available in a separate study.
Now we identify the relationships among the sectors. They are given by the
rows of the following table.
I.O. AGR PU M&M T&D CONS SERV
AGR X X X X
PU X X X X
M&M X X X X
T&D X X X X X
CONS X
SERV X X X X X
Given a certain sector i we ask: for any two sectors, h and k, to which sector are
more products from sector i allocated? The following treatments answer this
question for each sector.
110 8 Measuring Dependence Between Activities: Input–Output Application
Agriculture
The main crop in the Sudan is cotton. Cotton is exported and also allocated to the
manufacturing sector. Thus agriculture, transportation and distribution, and con-
struction do not receive a large amount of agricultural products. A new aggregate
is formed. (Note that only four sectors are considered under agriculture.)
8
< Agriculture
Aggregate
Transport and Distribution
ðAGGÞ :
Construction
As we pointed out above, the Sudan lacks adequate transportation. We aggregated
the two sectors which do not consume substantial quantities from agriculture, AGR
and T&D, because, although the main crop after cotton is wheat, the agricultural
sector allocates most of its output (i.e., wood) to construction. Transportation is
developed by means of loans from Arab oil countries and the World Bank. Thus, we
also aggregated agriculture and transportation to form a subaggregate.
AGR 0:9 0:0225
ð0:1Þ ð0:25Þ ¼
T&D 0:1 0:0025
Public Utilities
Construction
Construction only gives its products to services. Thus we associate the value 1
with services.
Services in the Sudan are very poor. We have assumed that the allocation of
service outputs to services, and to construction, are so negligible that these two
could be aggregated. We have
112 8 Measuring Dependence Between Activities: Input–Output Application
Construction
Aggregate ðAGGÞ
Services
Services
The matrix whose rows are the foregoing eigenvectors gives the distribution of
total intermediate outputs to the sectors. It is given by the following table.
Bibliography
9.1 Introduction
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 115
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_9, Ó Springer Science+Business Media New York 2012
116 9 Technological Choice in Less Developed Countries
In dealing with the problem of technology transfer, the AHP can be applied at
several levels. At a global level, one might be interested in predicting, in general
terms, the future course of technology transfers to the LDCs. A hierarchical rep-
resentation of this problem is depicted in Fig. 9.1. At the first level of the hier-
archy, we list the following actors who might be expected to play a role in
determining the future course of technology transfers: the governments of the
developed nations, the governments of the LDCs, ‘‘neutral’’ third parties such as
9.2 Applications to Technology Transfer 117
the many United Nations organizations involved in technology transfer issues, the
USSR and Eastern bloc nations, multinational corporations, the private sector as
represented by big and small businesses, and labor unions. Each actor pursues a set
of objectives such as technological leadership, trade, balance of payments,
employment generation, economic expansion, and so on, vis-à-vis technology
transfer. These objectives, spelled out in greater detail in Fig. 9.1, constitute the
second level of the hierarchy. At the third level we indicate the future scenarios: an
optimistic scenario that envisages increased flow of technology to the LDCs and
among world nations in general, a scenario that is merely the extension of the
status quo into the future, and a pessimistic scenario that foresees substantial
reduction in the overall volume of trade in technology. Increased flow of tech-
nology could occur if, for instance, the much debated code of conduct for tech-
nology transfer and the international economic order should become realities. The
pressures of the different actors and the strength with which they pursue current
118 9 Technological Choice in Less Developed Countries
policies could result in indefinite extension of the status quo, or the continued
impasse regarding technology transfer issues could deflect the LDCs toward pol-
icies of self-reliance and regional cooperation, thereby reducing the flow of
technology from the advanced to the less developed countries.
To estimate the likelihood of the three scenarios, we first make a pairwise
comparison of the relative influence of the different actors, which yields a set of
weights for this factor. Next the objectives of each actor are compared in pairwise
fashion in terms of their importance to the concerned actor. We pose questions
such as: ‘‘How important is objective O1 relative to objectives O2, O3, and so on,
for actor A1?’’ Having prioritized the objectives of each actor in this way, we
apply the weights of actor influence to these objectives to obtain a set of weighted
objectives. The process is continued by comparing the scenarios with respect to
their relative contribution to the achievement of each objective and by weighting
the scenario priorities by the weighted objectives. This results in an index of the
overall likelihood or importance of each scenario.
For the purpose of illustration, we have limited the hierarchy to only three
levels. The richness of the hierarchy can be extended as necessary by adding more
levels and more elements within each level. For instance, a level of actor policies
could be interposed between the objectives and scenarios, which might relate to
taxation, imports and exports, degree of encouragement given to foreign owner-
ship of capital, preferential treatments accorded to countries as sources of supply
of goods and technologies, and so on.
Let us now see how this approach can be used to enable planners of an LDC to
make technological choices. A survey of the literature suggests that the following
criteria are paramount in technology transfer:
1. Need: Technology has to be tailored to the needs of the country. Need can be
defined in terms of suitability and urgency. A technology might be considered
suitable if it is proven to meet similar needs in other contexts. Needs will be
influenced by overall national sectoral priorities.
2. Adaptability: It must be possible to adapt the technology to the local envi-
ronment. Two considerations determining adaptability are the ability and the
willingness to adapt. The lack of a sufficient science and technology base in
terms of the availability of skilled manpower, maintenance facilities, materials
in sufficient quantity and quality, presence of facilitating institutions and change
agents, and the like, affects the ability to adapt. The willingness to adapt
depends on the strength of custom, tradition, power relationships within the
society, and similar social, cultural, and political considerations.
3. Relative freedom from the risk of obsolescence: Here our attention is focused on
what may be termed ‘‘mass’’ technologies—technologies whose interface with
the eventual user is close and extensive. Many of these technologies are certain
to involve substantial investment. In the context of an LDC, the technology
transferred must be capable of being supported throughout its useful life; that is,
some insulation from the risk of obsolescence must be available. The most
modern technology may be difficult to adapt, while a currently obsolete
9.2 Applications to Technology Transfer 119
technology may at a later date leave the LDC without adequate support. Thus a
careful tradeoff has to be made as to the nature of the technology to acquire.
4. No undesirable second-order consequences: If the LDCs have any advantage as
late comers in the process of technological development, it is the opportunity to
profit from the lessons learned by the advanced countries as to the negative
effects of certain types of technology. Impact analysis must be an integral part
of any attempt to evaluate technologies for importation.
In general the needs for technology are influenced by sectoral priorities. Also,
each of the preliminary assessment criteria listed above will be emphasized dif-
ferently in the various economic sectors. For example, adaptability might be
stressed in agriculture, which is a more traditional sector than manufacturing in
most LDCs, whereas obsolescence risk might be underscored in manufacturing
(see [11]).
Sectoral priorities may be taken as ‘‘given’’ if they are based on planning
exercises done outside the context of technology assessment, a likely situation in
many LDCs. However, the sectoral priorities themselves can be hierarchically
determined. Two approaches are possible: a forward approach that starts with a
level of national objectives and works down toward the desired sectoral priorities,
and a backward approach that begins with constraints in both the internal and
external environment of the LDC and converges on the feasible sectoral priorities,
given the constraints (see [2]). Figure 9.2 illustrates the backward approach.
Returning to the problem of technology assessment (Fig. 9.3), we then have
sector priorities at level 1, assessment criteria at level 2, and candidate
120 9 Technological Choice in Less Developed Countries
In this example, which illustrates the approach described above, the sectoral prior-
ities are taken as given and assumed to have been determined exogenously. In the
context of a particular LDC, let us say the following are the sector priority weights:*
9.3 An Example: Technology Transfer Using the AHP 121
Agriculture 0.25
Mining and extractive industries 0.08
Health 0.12
Industry 0.15
Education and training 0.12
Communications 0.15
Transportation and distribution 0.06
Research and development 0.07
As already discussed, the assessment criteria do not play the same role in all the
sectors. They are prioritized within each sector in pairwise comparison matrices:
9.3.1.1 Agriculture
This matrix emphasizes that in the agricultural sector, the prime consideration for
introducing an imported technology is adaptability. Risk of obsolescence and undesirable
second-order consequences play relatively less important roles. These judgments, of
course, are those of a particular analyst who provided the pairwise ratings shown in the
matrix. Typically the entries of the matrix would be obtained after considerable debate and
ultimate convergence by consensus among the members concerned.
N A R S Weight
N 1 3 3 1 0.395
A 1/3 1 1 1/3 0.132
R 1/3 1 1 1 0.173
S 1 3 1 1 0.300
C.I. = 0.052
122 9 Technological Choice in Less Developed Countries
In the mining sector, in contrast to agriculture, need for the particular tech-
nology is felt to be the dominant criterion. ‘‘No undesirable second-order conse-
quences’’ ranks second. Adaptability and risk of obsolescence are relatively less
important, presumably owing to the state of development of this sector in the
particular LDC.
9.3.1.3 Health
N A R S Weight
N 1 1 3 7 0.402
A 1 1 3 7 0.402
R 1/3 1/3 1 3 0.143
S 1/7 1/7 1/3 1 0.057
C.I. = 0.002
The health needs of LDCs are widely different from those of the more advanced
nations. Hence need and adaptability dominate here.
9.3.1.4 Industry
N A R S Weights
N 1 1/3 1 1/3 0.132
A 3 1 3 1 0.395
R 1 1/3 1 1 0.173
S 3 1 1 1 0.300
C.I. = 0.050
N A R S Weight
N 1 5 9 3 0.587
A 1/5 1 5 1 0.172
R 1/9 1/5 1 1/5 0.044
S 1/3 1 5 1 0.196
C.I. = 0.033
9.3 An Example: Technology Transfer Using the AHP 123
The need for the particular type of educational technology is important, but
education that raises levels of aspiration can be dysfunctional if there are no
simultaneous increases in available opportunities. Hence a high score is obtained
on undesirable consequences of rapid education.
9.3.1.6 Communications
N A R S Weight
N 1 5 1/5 1 0.169
A 1/5 1 1/7 1 0.047
R 5 7 1 5 0.615
S 1 1 1/5 1 0.169
C.I. = 0.070
N A R S Weight
N 1 1/3 1/5 1 0.096
A 3 1 1/3 3 0.249
R 5 3 1 5 0.558
S 1 1/3 1/5 1 0.096
C.I. = 0.020
N A R S Weight
N 1 5 1 3 0.397
A 1/5 1 1/5 1 0.090
R 1 5 1 3 0.397
S 1/3 1 1/3 1 0.116
C.I. = 0.010
124 9 Technological Choice in Less Developed Countries
Now these sector-based criteria weights are adjusted by the sector priorities to
get the overall criteria weights for the particular LDC:*
Need 0.302
Ease of adaptability 0.314
No risk of obsolescence 0.230
No undesirable second-order consequences 0.154
9.3.2.1 Need
In the opinion of the analyst who provided the judgments in this matrix, rural
education (a long-term, low-risk project) and oil exploration (a short-to-medium
term, high-risk project) are the major national needs. Flood control and solar
energy are believed to be the next most important technologies from the point of
9.3 An Example: Technology Transfer Using the AHP 125
view of addressing national needs most directly. The consistency index of 0.18 is
rather high, even for a 7 9 7 judgment matrix. To some extent this happens
because of the broad types of comparison involved. In real life, one would be more
specific by clustering technologies so that they are more comparable.
Next, with respect to ease of adaptability, the relative standing of the candidate
technologies is as follows:
The weights obtained for each technology under each criterion are then
weighted by the criteria weights and summed to get the overall technology
weights:
126 9 Technological Choice in Less Developed Countries
The overall weighted priorities suggest that solar energy should receive max-
imum emphasis, followed by satellite television, flood control, oil exploration,
nuclear energy, computerized information bank, and coal gasification, in that
order. The high weight for solar energy is principally due to the high weight it
9.3 An Example: Technology Transfer Using the AHP 127
obtains with respect to ease of adaptability to a rural economy and the absence of
any conceivable second-order consequences associated with this technology.
Despite the high need for satellite television for rural education, it is perceived
by the analyst to be a potentially dysfunctional technology in the sense of dis-
tracting the masses rather than really educating them. Similarly the high need score
for oil exploration is counterbalanced by problems in adapting the technology to
local situations, obsolescence risk, and possible second-order consequences like
pollution of the coastal area. Nuclear energy gains on account of its relatively easy
adaptability and the fact that the technology is a rather stable one less prone to
obsolescence.
It is clear from this example that the AHP permits quantification of the relative
priorities of various technological options. The rankings obtained here reflect the
values, opinions, and judgments of a particular rater. But if the prioritization is
done in a group setting, then debate can be generated and judgments made on the
basis of consensus. The purpose of the example is to illustrate the application of
the AHP rather than to suggest any particular prioritization of technologies for any
particular LDC.
After the potentially most useful or desirable technologies are identified on the
basis of the four key criteria, conventional approaches such as cost/benefit analysis
can then be used to select from among the candidate technologies.
Figure 9.4 suggests a cost/benefit approach using the AHP. We construct
separate cost and benefit hierarchies and use the final cost and benefit weights to
derive cost/benefit ratios for the technologies under consideration. Thus we obtain
an ordering of the technologies in terms of their cost/benefit ratios.
Bibliography
1. Denison EF (1974) Accounting for United States economic growth, 1929–1969. Brookings
Institution, Washington
2. Emshoff JR, Saaty TL (1982) Applications of the analytic hierarchy process to long range
planning processes. Eur J Oper Res 10(2):131–143
3. Grilliches Z (1960) Hybrid corn and the economics of innovation. Science 29:275–280
4. Jones G (1971) The role of science and technology in developing countries. Oxford
University Press, Oxford
128 9 Technological Choice in Less Developed Countries
5. Kimenta J (1967) Economic theory and transfer of technology. In: Spenser DL, Woroniak A
(eds) The transfer of technology to developing countries. Praeger, New York
6. Mansfield EI (1961) Technical change and the rate of imitation. Econometrika 29:741–66
7. Organization for Economic Cooperation and Development (OECD) (1972) Analytical
Methods in Government Science Policy, Paris, 1972
8. Ramanujam V, Saaty TL (1981) Technological choices in less developed countries. Technol
Forecast Soc Choice 19:81–98
9. Rogers EM (1962) The diffusion of innovations. Free Press, New York
10. Rosenberg N (1972) Technological change and economic growth. Harper and Row, New
York
11. Saaty TL, Vargas LG (1979) Estimating technological coefficients by the analytic hierarchy
process. Socioeconomic Planning Sci 13:333–336
12. Schumacher EF (1975) Small is beautiful. Harper & Row, New York
13. Solo RA, Rogers EM (eds) (1971) Inducing technological change for economic growth and
development. Michigan State University Press, East Lansing
14. Solow RM (1957) Technical change and the aggregate production function. Rev Econ Stat
39:312–320
15. Spencer DL, Woroniak A (eds) (1967) The transfer of technology to developing countries.
Praeger, New York
16. Vernon R (1966) International investment and international trade in the product life cycle.
Quart J Econ 80:190–207
Chapter 10
Market Attractiveness
of Developing Countries
10.1 Introduction
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 129
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_10, Springer Science+Business Media New York 2012
130 10 Market Attractiveness of Developing Countries
the Strategic Management literature which is growing in importance with its own
approaches. Such names as ‘‘External Audit’’ and ‘‘External Factor Evaluation’’
[2] are being used to classify the environment. To model the discontinuity in the
present situation, the scenario approach has been used [7].
The authors’ company, Iritecna, has manufacturing activities in the general
engineering industry which include steel and non-ferrous metallurgy, logistic
activities for industrial plants and transport systems, and various involvements in
public utility infrastructures. It has locations in Latin America, North Africa,
Eastern Europe, the Middle and Far East. Its relationship to the world market is not
a simple transaction oriented buying-selling relationship. Rather, the relationship
with partners places emphasis on: (1) the quantity of money involved (about
$1,000,000 U.S.) and the significance of the length of a contract (preferably more
than 1 year); (2) the complexity of the rules of payment; and (3) the presence of
the company as a shareholder in some business activity.
The broad spectrum of industry segments in which the company operates as well
as its close involvement with the host country required by the particular way of
doing each type of business has suggested to the Planning Department the need for
an ongoing activity to monitor prevalent political and economic-financial trends in
the countries where it operates. In order of decreasing importance, the goals of this
undertaking are to create a process through a study and resulting report that:
• incorporates and correctly addresses the use of information, standard scale data,
ratings, and subjective judgments arising from different sources,
• supports and enhances the process of resource allocation for promotional and
commercial purposes,
• suggests new opportunities not well perceived by the company because of its
strong involvement in existing business activities,
• develops a final report containing a graded map with (cartographic) character-
istics, a kind of tactical board which portrays activity countries positioned
according to their socio-economic and political movements that influence the
macro-factors,
• makes possible learning about the future of the macroenvironment in which the
company operates,
• stimulates team discussion about the problem of where to go and when to go,
and utilizes expert knowledge and judgment for that purpose.
The last three goals clarify the intention to start a ‘‘macroenvironment modelling
activity’’ that enables involvement and utilization of the creativity and judgment of
both analysts and managers. Reports usually do not conform with an ongoing
modelling process, but must be examined item by item. The specialized literature [2]
suggests the use of weighted scoring methods to summarize and evaluate infor-
mation relating to ‘‘macroenvironmental analysis.’’ Multicriteria methodologies
such as Electre [9] specify a process to scale data and evaluate a final alternative to
follow for each decision. But these methods do not satisfy the need for a flexible
framework to represent an ongoing process of decision making. Multivariate sta-
tistical methods, used as geometric representations supporting Multiple Criteria
10.1 Introduction 131
Decision Making processes [3], were tested. It was difficult to interpret the final
outcome as a map of the environment because of its dependence on the measurable
statistical properties of the data rather than on the perception of the problem with its
social and political ramifications as they apply to each country.
The AHP [11] appeared to be well suited mainly for the richness of its frame
(hierarchical arrangement of the factors) which permits one to give evidence of how
the problem is perceived. It also allows for the use of intangible factors side by side
with tangible ones. It uses a simple procedure of paired comparisons to elicit
judgments from which it derives ratio scales. It then combines these ratio scales to
derive an overall ratio scale for the decision. Guided by a number of applications of
the AHP in the socio-economic planning field, it was decided to use it in making
this type of decision to: a) structure and map the problem in the economic and
political dimensions; b) interpret and use the results for decision purposes.
Second, the ‘‘strategic relevance’’ node could be decomposed into four sibling
nodes; underground resources, geographic position, historical importance, and
power sharing.
These nodes have not been included in the present version because our aim has
been to remain ‘‘simple’’ and relevant. We would also have encountered the dif-
ficulty of finding more specific data on them.
The third level of the model contains rating intensities by which the second level
criteria are measured: High, Medium High, Medium, Medium Low, and Low.
Where quantitative information is available, a relationship between rating
intensities and measurement data is used to make the evaluation (See Table 10.1).
All evaluations are ordered to give less desirable traits a lower rating. For example,
heavy turmoil, significant inflation and considerable risk for investment are all
given low ratings.
A country is evaluated for its ‘‘performance’’ with respect to each second level
criterion using the most appropriate rating grade to describe it. These results are
weighted and combined to yield weights with respect to the two major criteria. A
map of market potential displaying the synthesized results is the output sought.
10.3 Priorities
The AHP methodology requires priorities for each level of criteria and for each
rating intensity. These priorities are always determined in terms of their parent
node in the hierarchy. Relative value is derived from pairwise comparisons.
Judgments can be made directly with numerical scales or with a semantic scale
translated to numbers. For example, within the ‘‘Political’’ factors, the ‘‘Turmoil’’
subcriterion is strongly more important than ‘‘Strategic relevance’’ (scale value 5).
This strength of dominance is assigned for two reasons.
Table 10.1 Relation between rating grades and measurement data
10.3
Table 10.2 Pairwise comparison matrix components of the financial economic criteria
Risk of direct Current Inflation GDP per Growth rate Priorities
investment account/ rate person of GDP
GDP
Risk of direct 1 6 3 3 1 0.353
investment
Current 1 1/2 1/1 1/6 0.058
account/
GDP
Inflation rate 1 1 1/3 0.118
GDP pro 1 1/3 0.118
capite
Growth rate 1 0.353
of GDP
First present conditions are more relevant and pressing than long term and
uncertain evidence, and secondly, for ‘‘Turmoil’’ there is a significant amount of
information supplied by the specialized press and/or ratings institutes, while the
importance of ‘‘Strategic relevance’’ requires expert knowledge.
Evaluation of the elements under ‘‘Economic and financial’’ factors appears in
Table 10.2.
The resulting priorities show that:
• ‘‘Risk of Direct Investment’’ and ‘‘GDP Growth’’ are equally important and
together these two criteria dominate the remaining criteria.
• ‘‘Inflation Rate’’ and ‘‘GDP per Person’’ are next in priority and are significantly
less important than the previous two.
• ‘‘Balance Account/GDP’’ has the lowest priority.
These results mean that in order for a country to be attractive it should be in the
expansion phase and present low risk.
The AHP proved useful in creating intensity ratings for the next stage because
for most of the criteria there were no data for rating the alternatives and one had to
resort to the use of judgments. The resulting priorities are related to the relevance
given to the scale. For example, ‘‘Risk of Direct Investment’’ had no economic
indicator to evaluate the countries. Therefore, we had to create a scale of relative
intensities for the criteria and then used that to rate the countries. Here we did not
have the means to conduct and combine diverse analyses and to produce a
numerical outcome to be used in the model.
The pairwise comparison matrices for the five rating grades under each second
level criterion appear in Table 10.3a–g. The local priorities represent the relative
importance of a particular rating with respect to the parent criterion. It should be
noted that the differences in priorities decrease as one moves from the two highest
rating grades (High and Medium High) to the two lowest ones (Medium Low and
Low) for nearly all seven second level criteria. Exceptions are ‘‘Risk of Direct
Investment’’ and ‘‘Inflation Rate’’ for which it appears more correct that there
should be significant change from Medium High to Medium.
Table 10.3 Pairwise comparison matrix for the rating grades
10.3
(a) turmoil
High Medium high Medium Medium low Low Priorities
High 1 1.5 3 5 7 0.429
Priorities
Two axes were chosen to represent ‘‘Political’’ and ‘‘Economic and Financial’’
criteria. The five rating intensities are positioned along each axis in increasing
priority order. In synthesizing the local priorities established above, the rating
scales can be grouped into four segments with respect to each of the two major
criteria (see Table 10.4). The result is that the intersection of the two groupings
leads to a division of each segment of one into four subsegments.
Each of the segments and subsegments has a unique size. Here the AHP
analysis offers a different approach than the commonly used procedure called the
Boston Consulting Group matrix representation. Segment uniqueness is a feature
provided by the AHP to more accurately set the values of the five rating intensities
which is not possible in other methods known to us. Table 10.4 portrays an
interpretation of the four major segments in terms of overall market attractiveness.
• Eastern Europe
• Indian Subcontinent, and
• Far East
For historical reasons, the first four regions represent the reference market for
Italian companies exporting industrial plants and infrastructures. The last two are
included because in the last decade they have come to represent growing oppor-
tunities for Italian exports. There are also other reasons than historical ones that
necessitate tracing countries according to recent trends, to attract some of them
into influence zones, whose leaders are:
• U.S.A.;
• E.E.C.;
• Japan.
The countries included in this analysis (see Table 10.5) are the more important
ones in each region. They belong to the well known category of Developing or
Newly Industrialized Countries. Some relevant countries which one might expect
to find in Table 10.5, are missing mostly because of lack of complete data and/or
because they are in the process of splitting into sovereign states. South Africa and
Nigeria, which do not belong to the geographical regions mentioned above, are
included for their relevance to Italian companies. The country list reflects the fact
that the company originating the study is based in Italy and is dependent on its
government’s political and cultural relationship with other nations. Had the
company been French instead of Italian, the list could have been different,
including some countries in Central Africa.
The model evaluates each country under each of the seven second level criteria.
The evaluation is made by assigning the most appropriate rating intensity to a
country’s ‘‘performance’’ in each of the ‘‘performance’’ categories. The most
appropriate rating intensity is determined by the best available quantitative and
qualitative information. The study considers two time reference periods: 1990–
1991 and 1995–1997 (Table 10.6).
The sources of financial and economic information for each of the reference
periods are the following:
Period 1990–1991
• ‘‘Risk of Direct Investment’’ is based on rating data [1];
• ‘‘Current Account/GDP’’ is based on data from the ‘‘World Bank Annual Report’’;
• ‘‘Inflation Rate’’ is based on the annual report on risk forecast published by
Planning Review;
• ‘‘GDP per Person’’ was derived using data relating GDP and Population [12];
• ‘‘GDP Growth’’ is based on data from the annual report of Planning Review.
Period 1995–1997
• ‘‘Risk of Direct Investment’’ is based on rating data of the annual report of
Planning Review;
Table 10.5 Country ratings 1990–1991
10.4
Based on the rating data of the countries, a market attractiveness map is developed
in Table 10.4. The priority distribution of countries for the 1990–1991 time period
appears in Table 10.7. Table 10.8 represents the change in priorities of the
countries from the first to the second time period according to region. Table 10.9
shows the priority distribution of the countries for the 1995–1997 time period.
144 10 Market Attractiveness of Developing Countries
Table 10.8 Shifting in countries map position reference time period 1990–1991
troubles than Eastern Europe. Vietnam is considered by the specialized press [5] to
be very interesting because of its skilled, disciplined and cheap labor force.
• Below the main diagonal there is a set of countries (Indonesia, Colombia, Paki-
stan, Philippines, and Nigeria) that have a medium economic-financial position
but a poor political position. Within this group there are two countries, Columbia
and Pakistan, which because of drug trade and investment in the nuclear sector
may become a serious cause of instability in the North–South relationship.
• At the end of the main diagonal there are three Latin American countries
(Ecuador, Peru, and Bolivia) which even without a negative environment in
Latin America do not show clear signs of progress.
10.6 Conclusion
The observations made above on the countries‘ positions, show the effectiveness of
the AHP in discriminating between different situations. But there also remain a
number of countries in the center of the map that may need further separation,
perhaps by expanding the second level criteria. Such work is in progress. It
requires user interface modification of software implementation, combining the
hierarchic composition and its resulting map. The decision maker is given the
opportunity to add criteria, modify their weights and note the resulting map. This
modification could allow a fine tuning of the model which would facilitate sen-
sitivity analysis. Sensitivity analysis is particularly needed to deal with political
change. It is believed that in the current environment, ‘‘Turmoil’’ is more
important than ‘‘Strategic relevance’’. It seems important to explore changes in
countries’ positions by shifting emphasis from the present situation (‘‘Turmoil’’) to
the long term trend (‘‘Strategic Relevance’’). Similarly, by scrutinizing the eco-
nomic-financial side of the model, it appears promising to assign higher priority to
indicators of richness and to the competitive position (GDP per Person, Balance
Account/GDP) to obtain greater resolution at the center of the map.
The outcome of this study has been well received within the company with
interest being shown by managers through their participation. They began with the
final map and worked backwards to the structure of the model making suggestions
for further improvements. The map provides evidence for the validity of the
impact of macroenvironmental changes on international business and it is a valid
support for establishing a connection between scanning the macroenvironment and
perceptions of strategic issues [8].
Bibliography
3. Marescal B, Brans JP (1991) Geometrical representation for MCDM. Eur J Oper Res 54:6
4. Le Monde Diplomatique (1993) no. 486, March 1993
5. Le Monde Diplomatique (1993) no. 469, April 1993
6. Le Monde Diplomatique (1993) no. 169, April 1993
7. aa.vv., Monografic issuesissues on Scenarios (1992) Planning Review, March–April, April
June 1992
8. Morecroft JDW (1992) Executive knowledge models and learning. Eur J Oper Res 56:1
9. Roy B (1990) The outranking approach and the foundations of Electre methods. In: Bana e
Costa CA (ed) Readings in multiple criteria decision aid. Springer, New York
10. Rullani E (1992) due problemi irrisolti per il capialsimo industriale negli anni novanta.
Economia e Politica Industriale 74:76
11. Saaty TL (1990) Multicriteria Decision Making: The Analytic Hierarchy Process. RWS
Publications, 4922 Ellsworth Avenue, Pittsburgh
12. World Bank Annual Report (1993)
Chapter 11
An Analytic Hierarchy Process Based
Approach to the Design and Evaluation
of a Marketing Driven Business
and Corporate Strategy
11.1 Introduction
A marketing driven strategic business plan requires the generation, evaluation and
choice of eight interrelated components:
(1) Mission
The mission should offer an explicit, visionary, and unique direction for the
entire planning process. It offers a first cut at the determination of the business
boundaries of the firm. It also serves to mobilize the firm to act and differentiate it
from the others.
(2) Planning horizon
Planning should accommodate both the short- and long-term needs of the firm.
Explicit tradeoffs between the two time horizons should be identified.
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 149
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_11, Springer Science+Business Media New York 2012
150 11 An Analytic Hierarchy Process Based Approach to the Design and Evaluation
strength in the segment (also a composite score which includes measures of the
firm’s competitive strengths such as technology, distribution, its market share etc.),
and (c) synergy—a dimension recognizing the interdependency among activities
which, if not included explicitly as part of the strength dimension, should be used
as a separate dimension.
(6) Market/product portfolio
Identification of the current, potential, and desired market/product portfolio is
the focal point of the planning process. The desired market/product portfolio
includes current markets/products and new ones developed either internally or
externally (via marketing and advertising). The market/product portfolio is the
core marketing dimension of the process. It is this focus on the selection of a
portfolio of market segments which differentiates a marketing driven from a non-
marketing oriented business or corporate strategy. The portfolio of segments and
their associated products define the business boundaries of the firm and to the
extent that the firm employs a global perspective, it would also incorporate the
portfolio of countries by mode of operation. This step often involves three inter-
related processes: (a) identification of new market segments (for target product by
country); (b) evaluation of the current and new segments on the criteria and
objectives specified in points (4) and (5); and (c) the generation of and evaluation
of a portfolio of products/segments which reflects portfolio considerations such as
diversification versus focus or, in the international context, the desired levels of
integration and coordination of product/segments across countries.
(7) Strategic options
Creative options should be generated to meet the needs of the market segments
and offer the firm a unique competitive advantage. These options should identify
the major leverage points the firm has (e.g. research and development, manufac-
turing, marketing, finance etc.) and the strategic thrust most likely to meet three
interrelated criteria: (a) meeting the segment needs; (b) differentiating the firm
from its competitors; and (c) accomplishing the firm’s own objectives. The
identification of strategic options should also include determining whether the
strategy can be accomplished internally or requires a merger, acquisition, or other
forms of strategic alliances. Since the generation and evaluation of strategic
options is the primary outcome of any strategy process, it is especially important to
assure that before making the final selection, the participating managers examine
whether better strategies can be developed by increasing, reducing, eliminating, or
adding activities (products, segments, countries, distribution systems etc.) or
reallocating resources among the various options.
(8) Functional and resource requirements
Once the strategic options are identified, the functional requirements (from each
operating department) for meeting the needs of the strategic options should be
identified and evaluated.
152 11 An Analytic Hierarchy Process Based Approach to the Design and Evaluation
Strategic SO1 SO 2 SO 3 SO 4 SO 5
Options (SO)
to Meet the
Needs of the
Segments:
The selected methodology for implementing this approach is the AHP [2, 3].
Following this approach, the entire process can be summarized in a single hier-
archy in which the lowest three levels present the desired business and corporate
strategies. The evaluation of the strategies must reflect the relative importance of
all the considerations at the higher levels of the hierarchy. Thus, in the hierarchy
illustrated in Fig. 11.1, the priorities for allocating resources among the functional
requirements are weighted (sum of the weights of each row is always 100) by their
contribution to the achievement of the strategic options (new acquisitions, new
markets to enter etc.). These options are in turn evaluated with respect to their
ability to achieve the desired product/market portfolio of the firm. The selected
portfolio is in turn evaluated on its market attractiveness, business strengths, and
synergy. These composite criteria are then weighted according to their importance
to the achievement of the objectives of the firm, which in turn are weighted on
their importance under (a) the expected scenarios and their likelihood of occur-
rence in the short and long term and (b) management tradeoff between short and
long term horizons in accomplishing its overall mission.
An AHP-based approach for the development of a marketing driven business
and corporate strategy provides:
11.3 An AHP Formulation of a Marketing Driven Business and Corporate Strategy 153
11.4 Applications
The marketing driven planning process described in the previous section and
outlined in the hierarchy of Fig. 11.1 has been applied in a number of cases
ranging from the selection of a portfolio of segments and products to the design of
an overall SBU and corporate strategy. These applications involved a number of
major U.S., Japanese and Latin American firms. Given the sensitive nature of some
of the applications, the discussion in this section will focus on the lessons learned
from them, in particular; key conclusions from the applications, modifications
made to the basic framework of Fig. 11.1, areas requiring development.
154 11 An Analytic Hierarchy Process Based Approach to the Design and Evaluation
Reflecting on the experience gained from six applications of the process described
in Fig. 11.1, the following conclusions can be reached:
• Top management teams have no trouble using the AHP
• Having a structure such as the one proposed in Fig. 11.1 helps speed and
facilitate the planning process
• The process as outlined in Fig. 11.1 can be used both at the SBU and corporate
levels
• The development of a mission statement is often a difficult task. The business
definition component of the mission is often revised after completing a first run
through the hierarchy.
• Regardless of management’s initial evaluation of the tradeoff between short and
long term, it is helpful to present them with the results of a sensitivity analysis
that encompasses the entire range of options from 90/10 short versus long term
to 10/90 short versus long term.
• The construction of scenarios is a task requiring significant staff input, espe-
cially if this step is used as a basis for situation analysis involving both internal
and external forces. New competitive entries, dramatic technological develop-
ments, regulatory changes, and significant shifts in the demographic and life-
style characteristics of the target markets offer useful starting points for the
construction of key environmental scenarios.
• All the applications to date have involved multiple objectives. The most com-
mon ones are profit and sales levels and growth and reduction of downside risk.
Other objectives tend to reflect idiosyncratic characteristics of management. In
all cases the relative importance of specific objectives varies depending on the
expected scenario. Typical relationships involved increased importance of
profits as the scenario became less favorable.
• Most managers have difficulty directly evaluating market segments and busi-
nesses as they relate to their firm’s objectives. The use of market attracti-
veness and business strength as intermediary composite criteria does not
simplify the process. There is no agreement, however, across the various firms
as to the specific components of each of the two criteria. The GE/McKenzie
conceptualization offers a good starting point but is often modified by
management. Managers also tend to vary in their evaluation of the relative
importance of various components of these criteria as well as the relative
importance of market attractiveness and business strength—some prefer to
focus on attractive markets, while others prefer markets in which they have
specific strengths.
• In most applications, synergy was viewed as a separate and significant criterion.
Effective evaluation of positive and negative synergy among the segments and
other portfolio operations requires, however, the identification of specific
dimensions of synergy (e.g. distribution, manufacturing, procurement, etc.).
11.4 Applications 155
• The focus on the market segments to be served and their associated products is
initially difficult, especially at the corporate level, but once explored, it greatly
simplifies and focuses the remaining task of developing creative strategies to
satisfy the needs of and benefits sought by the selected segments.
• The evaluation of the current and expanded portfolio of segments and products
(on the selected criteria) is a relatively straightforward task. It is easily sum-
marized in matrix form as segment/products by criteria, or on a market at-
tractiveness business strength chart. It is much harder, however, to use the
results of this evaluation as a guideline for the selection of innovative portfolio
strategies (e.g. diversification, focus acquisitions etc.). Separate analysis should
be undertaken focusing on the generation of portfolio strategies presented as a
new level in the hierarchy. These portfolio operations reflect the results of the
previous analysis and other considerations and should be evaluated on the
objectives of the SBU or the firm.
• It has been helpful in all cases to augment the domestic portfolio of market
segments and products with a global perspective focusing on segments by
products by country by mode of operation. This is not an easy task, and lack of
comparable information across countries has been one of the major obstacles.
Such an implementation requires heavy reliance on the subjective judgments of
experts.
• In a number of cases, management found it helpful to supplement the basic
hierarchy with a direct evaluation of the segments and their associated products
(and countries by mode of operation) on the requirements for success versus the
firm’s expected strengths under each of the expected scenarios.
• The generation of strategic options has greatly benefitted from the use of
analogies and other approaches for enhancing creativity [4, Chap. 9].
• The focus on market segments and their associated strategic options helps in
integrating the various functions (e.g. R&D, manufacturing, finance, marketing
etc.) in a coherent, focused direction.
• All the participants in the various applications found the sensitivity analysis to
be of great value and a useful input to the revision of previous decisions.
SCENARIO: S1 S2 S3 S4 S1 S2 S3 S4
OBJECTIVES: O1 O2 O3 O1 O2 O3
It provides the framework for a series of iterations between the left- and right-
hand-side hierarchies. The process starts by identifying the ‘‘best’’ strategy for the
firm (say, segment/positioning A and strategic thrust 1). This strategy is now
introduced as part of the scenario facing the key competitor (say, scenario 1) and
the competitor’s best strategy, reflecting our strategy as part of its environment, is
now assessed. This strategy (segment/positioning D and strategy thrust 2) is now
considered part of the scenario of our firm, and the previous strategy and other
strategic options are examined against it to assure that the selected strategy is the
best one. The series of iterations can continue until a ‘‘quasi optimal’’ strategy is
found. In a number of applications of this procedure, about five iterations were
required to select the best strategy.
(b) Developing supplemental hierarchies. In a number of cases, it was found
that the basic hierarchy presented in Fig. 11.1 had to be supplemented with a more
specific hierarchy for the completion of specific tasks. For example, in deciding on
an R&D/technology licensing strategy, it was found that the top six levels of the
hierarchy in Fig. 11.1 provide the basic structure but that the two new and
interrelated hierarchies with additional Levels 7 and 8 should be developed. The
product concepts most appropriate to reach each segment are placed in Level 7,
and R&D project or technology licensing and acquisition options required to
develop the product concept in Level 8. A second hierarchy was then developed in
which the various R&D projects and technology, licensing and acquisition options
(Level 8) were evaluated on a new set of criteria incorporating likelihood of
success, expected cost and completion date, and synergy among the projects (as a
new Level 7).
11.4 Applications 157
The AHP using the market driven hierarchy outlined in Fig. 11.1 works! It can,
however, be further improved by:
• Simplifying the data collection task by reducing the number of required
judgments.
• Integrating diverse data collection procedures such as the basic reciprocal
matrix using a 9-point scale with 100 points constant sum allocation or ranking
for the evaluation of a large number of options.
• Integrating other data sources, especially market response functions and envi-
ronmental scanning and forecasting, with management subjective judgments.
• Incorporating management uncertainty in various judgments.
• Linking the resulting priorities with optimization programs leading to optimal
allocation of resources.
11.5 Conclusions
The AHP offers a unique and valuable method for the generation and evaluation of
marketing driven business and corporate strategy. The basic hierarchy presented in
Fig. 11.1 which focuses on a portfolio of target segments, assures that the planning
process will be market driven.
The process has been successfully implemented in six diverse cases. The
experience with these applications reinforces the favorable results obtained in
many other applications conducted by Saaty and his colleagues. These results
suggest that the process is easily implementable and offers a relatively quick and
simple approach to business and corporate planning process.
The applications do support the need for further refinement of the data col-
lection part of the process, its integration with other data sources and analytical
procedures.
Another interesting future development is the linkage of AHP to a series of
expert systems. Expert systems could facilitate the accumulation and synthesis
of knowledge, in the discipline itself and in the participating firms, particularly
for such processes as portfolio analysis and strategy. Such development can be
modeled for advertising messages [1]. Yet, even without such a development
the AHP greatly facilitates a marketing driven planning process and encourages
the generation of creative solutions and their rigorous evaluation.
158 11 An Analytic Hierarchy Process Based Approach to the Design and Evaluation
Bibliography
1. Rangaswamy A, Burke R, Wind Y, Eliashberg J (1986) Expert systems for marketing, wharton
school working paper. University of Pennsylvania, PA
2. Saaty TL (1977) A scaling method for priorities in hierarchical structures. J Math Psychol
15(3):234–281
3. Saaty TL (1980) The analytical hierarchy process. McGraw-Hill, New York
4. Wind Y (1983) Product policy: concepts, methods, and strategies. Addison-Wesley, Reading
5. Wind Y, Saaty T (1980) Marketing applications of the analytical hierarchy process. Manage
Sci 26(7):641–658
6. Wind Y (1987) An analytic hierarchy process based approach to the design and evaluation of a
marketing driven business and corporate strategy. Math Model 9(3–5):285–291
Chapter 12
New Product Pricing Strategy
12.1 Introduction
This chapter presents the development and application of a model for effective
decision making in establishing strategies for the pricing of new products. The
model developed evaluates all important criteria that need to be considered for
the successful implementation of new products in the market. The formulation of
the model was tailored for a specific new software product with unique marketing
considerations in a well-defined, segmented market. The process used can easily
be extended to include other products, provided that model changes and other
appropriate parameters realistically describe the problem being analyzed. Such
information needs to be established on a case by case basis by the user, for proper
validation of the model.
The model has been applied to a case study involving an actual situation related
to the introduction of a large new software product in the market place. The
analysis performed considered potential sales and customer benefits covering both
pessimistic and optimistic scenarios in an attempt to bound the impacts of this new
product, and to support the decision on the best pricing strategy for the intro-
duction of this product in the market. The product being marketed is an advanced
high technology software program. Potential buyers will be identified by seg-
menting the market into groups. For example, market research indicates that the
client base consists of electric utilities operating nuclear power plants. Due to the
homogeneous nature of the client base, and as a result of extensive government
regulations, multiple market mixes are not appropriate.
The product was developed to meet the customer’s desire to accurately monitor
reactor cores by providing advanced technology. The act of tailoring the product to
the buyer’s desires is consistent with demand-side marketing. Substitute products
are not currently available to potential buyers of this new software. This is
essential for changing the technological advantage currently present into a lasting
commercial advantage.
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 159
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_12, Springer Science+Business Media New York 2012
160 12 New Product Pricing Strategy
The process of positioning allows the supplier to develop a broad range of uses
or applications for the new product. It is essential that every effort be made to
identify as many applications as possible, both current and future, for this software
product’s promotional campaign. As the product was designed to meet the demand
for sophisticated reactor core monitoring capabilities, one of the primary sources
for obtaining information concerning applications of the product will be the cus-
tomer. The installation of a prototype at a customer’s site has provided the
opportunity to collect feedback concerning the customer’s need for product
modifications, enhancements as well as new applications. In addition, the foun-
dation was laid for increasing customer loyalty to the product. Significant data was
collected during the on-site demonstration of the software. This provided valuable
marketing, research and development insight concerning potential applications in
response to future customer needs.
The promotion of a new product encompasses advertising, personal selling,
sales promotion and public relations. Brochures, pamphlets, catalogs and other
print media, announcing the product, are produced by the in-house marketing
department. The documents are periodically updated to reflect the new applica-
tions and enhancements that are identified as a result of marketing research and
customer feedback. Supplies are given to the nuclear fuel vendor to be distributed
during on-site visits to the plants. In addition, plans have been made to produce a
video storyboard to be used in conjunction with more formal product presentations
to senior management and regulatory agencies. Although the product will be sold
by the nuclear fuel vendor, the marketing department provides follow-up assis-
tance to potential buyers, attends trade shows and provides literature, and operated
a speakers bureau service. The strategic marketing plan targets a very narrowly
defined market segment. Marketing is essential due to the constant rate of change
in a high-technology market and to ensure that potential buyers are fully informed
of the software’s capabilities. Following several years of development, the pro-
gram is now available for customers.
Another dimension which is important in the development of a marketing
strategy is product pricing. Classical economists in the eighteenth and nineteenth
century viewed price as responding solely to market conditions. They believed that
prices were a function of supply and demand alone. The economic system was
supposed to automatically set the price of all products.
Just over 40 years ago, an economist, Gardiner C. Means, defined the term
‘‘administered prices’’ to describe goods which had rigid prices and did not fluc-
tuate with a variation in demand. Today, ‘‘administered prices’’ are formulated in
corporate offices of most companies as a matter of operating policy, financial
planning, and marketing strategy. Nowhere is an ‘‘administered price’’ more a
factor than with a new product. For a new product, there is a range of alternative
pricing policies which can be adopted. Each alternative policy suits specific cir-
cumstances. The pricing policy alternatives relevant to our product are skimming,
penetration, and strategic.
Skimming achieves the maximum profit in the shortest possible time by charging
the highest price the market will bear. Penetration achieves the maximum market
12.1 Introduction 161
All the parameters included in the model developed have been tailored for the
specific case of a new computer program software that provides sophisticated
capabilities for the monitoring of nuclear reactor cores. As such, this product has
targeted specific customers, which required unique considerations of the market
and distribution channels. However, the considerations made in the model and the
hierarchy can easily be extended for different situations, if appropriate criteria are
identified and specific judgments are used in the analysis.
Several different scenarios were evaluated to consider the uncertainties in the
planning scheme over the next 5 years; the time period of acceptance for this new
product in the market.
The product considered is a sophisticated computer program for the support of
operation in Nuclear Power Plants, with the capability to display margin to safety
limits. However, it is not part of the plant automatic protection systems. This
system provides information for operators, site engineers and designers on reactor
core performance. The software operates on engineering workstations and inte-
grates several independent functions at the power plant site.
This software has been under development for 4 years, and has resulted in a
large computer program of approximately 70,000 lines of code. Considering the
technical complexity and the interactive nature of its user interface, an extensive
site demonstration was conducted during the past year, with the installation of the
product at a host site for use in actual situations. This demonstration program was
very successful, leading the Beta-Test customer to acquire the software for all its
plants. It provided significant and valuable data on product modifications and
enhancements that address future customer needs. This product introduces leading
technologies in both software and hardware, with the potential of altering pro-
cesses that can limit its acceptance.
The market for this software consists primarily of Electric Utilities that operate
nuclear Power Plants. In addition, the product is more applicable to a specific class
of reactors. This clearly identifies the potential participants in the market, thereby
162 12 New Product Pricing Strategy
bounding clearly the potential demand for the product. The finite number of
consumers involved and the highly technical nature of its application determine
that an extremely focused marketing strategy should be employed.
At present there are no competitors who can offer products of similar tech-
nology and capabilities. However, there are strong indications that software is
being developed that may match this product’s functions, but such developments
are at least 2–3 years from any practical implementation.
Since there is product exclusivity in the market, the model and economical
considerations made in this paper are representative of a monopoly. However,
initial market analysis has indicated that price is still an important parameter due to
the approval of customers Operating and maintenance expenses by Public Utilities
Commissions.
No specific considerations are required on distribution channels for the product
as marketing, delivery and support will be provided by the nuclear fuel vendor.
Commercial and technical relationships with potential customers are well estab-
lished. These relationships provide an additional competitive advantage as com-
petitors cannot offer the same level of support nor do they have the same leverage
in packaging their product with other large contracts.
This software also provides other benefits to the supplier. In addition to
sales, it strengthens existing supplier-customer relationships. The implementa-
tion of this product will act as a small barrier for competitors since they will
be required to invest substantially supporting the system. It has a significant
impact on the technical processes involved as it introduces new approaches that
alter the technical interface between supplier–customer. These changes provide
opportunities for additional businesses and differentiation in the market place,
which are more significant than the technological leadership introduced by the
product.
In sum, the objective of this model is:
To determine the best pricing strategy for a new software product that takes into account
relevant considerations of market share and return on investment, and accounts for the
overall uncertainties associated with product acceptance projections.
The hierarchy designed for this model (see Fig. 12.1) evaluates the pricing
strategies according to defined limits of relevant business criteria. Each needs to be
considered for the particular problem in question, as well as for different scenarios
that reflect uncertainties in the planning period.
The Pricing Model described here has been established to guide pricing decisions
of a new software product. It includes alternatives on time horizon uncertainties
about the market behavior over the next 5 years. Three scenarios have been
included in the hierarchy:
12.2 The Analytic Hierarchy Model 163
• The Optimist Market Scenario is the result of a set of circumstances that lead to
the capture of the entire market, i.e., maximum market share,
• The Pessimistic Scenario establishes a set of potential situations that lead to a
minimum market share, and
• The Best Estimate Scenario is the middle of the road, in which a realistic market
potential will be possible for this product. Preliminary market research has
indicated that this approach is the most realistic in the near term until the
technology is fully approved by the licensing authorities.
These scenarios are described in detail in Table 12.1.
The potential pricing alternatives need to be evaluated according to the set of
criteria relevant to the business parameters established for this product.
12.2.2 Criteria
Five criteria that influence the product price were identified. In the Analytic
Hierarchy Process, each is compared against the others to determine their degree
of influence on the goal:
Market—The potential market share achievable is an important parameter in
defining the product price, as product development costs make a large contribution
to the total product costs, and therefore, the profit. In addition, market share is
important from the perspective of potential additional business through product
upgrades and preempting the possibility of competitor entry into the market.
Supplier ROI—The Return on Investment is another important business
consideration from the suppliers point of view, and product sales should meet
164 12 New Product Pricing Strategy
The pricing strategies developed for this model address the classical approach of
marketing high technology products, and the seller’s objectives (Fig. 12.2):
Skimming Strategy—This strategy suggests to demand a premium for the
product, which will result in a limited number of sales but with high overall ROI
for the seller. This strategy could be justified in that the product provides signif-
icant benefits for customers, and the technology lead over competitors. However, it
is the alternative with maximum risk, in that a high price makes the justification of
the product purchase more difficult. There are several examples in product
implementation that have adopted this approach with disastrous results.
Penetration Strategy—The Penetration Strategy proposes a pricing level that
provides certain incentives for more customers to purchase the software, and at the
same time, provide a good return to the supplier. This is a middle of the road
approach that attempts to balance benefits for both the supplier and customer while
ensuring a broad market for the seller.
166 12 New Product Pricing Strategy
In this section, we discuss the results obtained with the model developed in Sect.
12.2. In addition to the reference model result, a sensitivity study was conducted to
establish the reasonableness of the reference case and to identify the conditional
parameters that would lead to a different course of action. The results of the
sensitivity analysis provide insight into trigger points that could be established and
monitored during the implementation of this product, and would indicate the need
for reassessment of the initial pricing position.
The model was input into the ‘‘Expert Choice’’ AHP-based software program and
run to define the base case pricing strategy for the new product introduction. As
discussed in the previous section, fixed and variable cost structures and the
required supplier return on investment were provided by corporate instruction.
Pairwise comparison values of the elements within the model were designed by
consensus within the group.
Based on the parameters discussed and the synthesis of the relevant criteria we
conclude:
Alternatives Priorities
Skimming 10.8%
Penetration 42.8%
Strategic 46.4%
and hence, the overall best pricing strategy for the new product is the Strategic
alternative.
The strategic pricing approach incorporates the lowest price of the three strategies
and facilitates product acceptance. In this case study, significant importance was
placed on market share and provision for future profits through upgrades to the
software; a long-term approach to economic performance of the product is used.
Supplier’s return on investment and delivery costs were the secondary concerns and
customer return on investment was the least significant concern.
12.4 Sensitivity Analysis 167
Sensitivity tests of the model to variations in the market demand were run to aid in the
prediction of future changes in pricing strategy. In addition, the model’s sensitivity to
alternations in corporate strategy with respect to return on investment was tested.
Figure 12.3 illustrates the effect of the variations in the market strength on the
pricing strategy employed in the reference case. For example, in a pessimistic market
environment, strategic pricing is the most desirable approach. In a weak market
environment, the supplier must be concerned with recouping sunk costs to a large
extent and realize that a weak performance will bear on the viability of future projects.
As the estimate of market strength increases, a continued low price will provide
maximization of market share and thus eliminate barriers to product acceptance. The
strategic approach is the best alternative for this reason as well as the firm’s con-
tinued emphasis on its long-term objective of additional business due to upgrades.
In an optimistic market environment, the penetration approach to pricing
becomes increasingly more desirable than the strategic approach. More emphasis
will be placed on the suppliers return on investment.
Market Uncertainties—Variations in the demand for the new product results in
a shift in the pricing strategy. As presented in Fig. 12.4, an expanded (optimistic)
market would strongly reinforce a continuation of strategic pricing. A shift in
market demand to one worse than predicted also dictates a shift in pricing policy.
168 12 New Product Pricing Strategy
The model is sensitive to reduced demand and indicates that a ‘‘soft’’ market
would signal a switch to penetration pricing.
Corporate Objectives—Additional sensitivity tests were also run on this model to
evaluate the effect of variations of corporate strategy on the pricing strategy. A shift in
emphasis from long-run planning to short-run returns was tested. With the emphasis on
short-run profits, the optimal strategy in all environmental scenarios, as defined by the
model, is penetration strategy (Fig. 12.4). A comparison of short-run vs. long-run
pricing strategies in a best estimate environment is presented in Fig. 12.5.
12.5 Conclusions
The decision as to how to price any product is a complex one. This process must be
approached with a systematic methodology that accounts for all important criteria.
The criteria used will vary with the product. In this study, the product being priced
is an advanced software program that is used to monitor nuclear reactor cores.
The market mix is limited due to the nature of the possible customers. This market
can be thought of as a monopoly in that there is only one viable producer of this
product.
12.5 Conclusions 169
A review of the market showed that there are three possible pricing strategies
for this product. These include:
• Skimming—Maximum profit in the shortest amount of time.
• Penetration—Low price, high market share.
• Strategic—Extreme variation of penetration. Lowest price for the greatest
market share.
It was from these three strategies that a best case must be selected. The results
demonstrated that in the long-run, the strategic pricing approach would be best for
this product. However, sensitivity analysis of the model showed that a switch to
penetration pricing may be necessary if the market were to turn soft. In addition, if
the emphasis of the company was to shift to short run profits, penetration pricing
would become the strategy of choice regardless of the market strength.
A final conclusion is that the results of the sensitivity analysis could be further
studied to precisely determine trigger points where a change in pricing strategy
would be necessary. These trigger points would be the result of any change in
company direction that altered the assessment of the model criteria or a significant
change in the market conditions.
Chapter 13
Incorporating Expert Judgment
in Economic Forecasts: The Case
of the U.S. Economy in 1992
13.1 Introduction
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 171
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_13, Springer Science+Business Media New York 2012
172 13 Incorporating Expert Judgment in Economic Forecasts
Studies of ‘‘ex ante’’ forecasts produced by the builders of major models using
add factors suggest that these forecasts have been more accurate than the ‘‘ex post’’
forecasts produced by the models themselves, even when the same add factors
were employed. Fair ([3], p. 263) wrote:
In other words, the use of actual rather than guessed values of the exogenous variables
decreased the accuracy of the forecasts… This conclusion is consistent with the view that
the add factors are (in a loose sense) more important than the model in determining the ex
ante forecasts…
All of this suggests that macroeconomic model builders/forecasters are well aware
of the limitations of their underlying models and the need to incorporate subjective
judgments, especially in the face of structural shifts in the models. However, these
judgmental adjustments are necessarily non-systematic and ad hoc in nature.
Rather than further critiquing this approach, however, we instead provide an
alternate which is both systematic and consistent in its ability to capture the impact
of structural changes. While we have not illustrated this alternative by adapting a
formal macroeconomic forecasting model, we do employ an eclectic conceptual
framework grounded in modern macroeconomics. Our alternative, moreover,
could also be readily employed to enrich forecasting exercises based on formal
models (e.g. generating add factors more systematically and consistently; adjusting
the values of exogenous variables). In this respect, the two forecasting approaches
can be seen to converge quite compatibly.
1
An alternative measure of U.S. aggregate economic activity is now used by the U.S.
Department of Commerce: namely, gross domestic product (GDP), which nets out international
factor payments and receipts. We will continue to cite GNP patterns in this chapter for historical
purposes. Currently, real GNP is running slightly higher than real GDP.
13.3 The Setting: A Sluggish Recovery/Structural Change 173
rate of the expansion through the second quarter of 1993 is now stronger but still
quite modest, about 2.7%. Thus, even if we agree with the NBER’s dating of the
most recent cyclical trough, the expansion’s initial phase must be judged the
weakest in recent history. Table 13.1 provides evidence for the first six quarters of
the three previous expansions side by side with the most current expansion (using
the NBER dates for the troughs):
Survey of Current Business, US Department of Commerce, various issues,
NBER dates are used to identify troughs
Moreover, other general economic indicators have performed sluggishly. Most
importantly, total employment failed to grow appreciably during this putative
expansion, unlike the three previous expansions.
It needs to be emphasized that much of this information was unavailable to us
when the authors convened in late-December 1991/early January 1992 to engage in
an exercise aimed at forecasting the trough of the current cycle. At that point, we
did not subscribe to the notion that a less than 1.0% rate of real GNP growth from
the first quarter 1991 to the third quarter 1991 (the latest data then available)
signified an economic expansion, especially when coupled with other economic
data being reported at that time. When we reconvened in May 1992 to review our
previous assessment and to engage in an exercise aimed at forecasting the strength
of the eventual recovery, the annual rate of real GNP growth through the fourth
quarter of 1991 (the latest then available) had apparently actually weakened mar-
ginally. The rate of real GNP growth between the last quarter of 1991 and the first
quarter of 1992 subsequently rose to 3.5% on an annual basis, only to decelerate
again in the second quarter to something of the order of 0.7%. The growth rate then
grew substantially in the third quarter of 1992, perhaps signalling the onset of a
stronger recovery. Again, this information was obviously not available to us in early
May of 1992, though other negative economic portents certainly were.
It is also instructive to compare the forecasts produced in this chapter with those
prepared at the time by professional economic forecasters. Blue Chip Economic
Indicators is a monthly publication which reports the consensus forecast of 50
major economic forecasters, including teams at universities, banks, corporations,
forecast specialist firms, and professional and credit evaluation institutions.
In December of 1991, the Blue Chip CONSENSUS forecast for real GNP growth
in 1992 was 2.2%. This represented a reduction of two-tenths of a percentage point
from that projected in the previous month. However, the CONSENSUS reported
that: ‘‘…, the seeds of growth have been planted and next spring should see a
sustained economic recovery begin to sprout.’’ Further, the CONSENSUS in March
of 1992 indicated that the economy is ‘‘…perking up…’’ and suggests ‘‘…guarded
optimism.’’ for the expected recovery. More importantly, the CONSENSUS shifted
174 13 Incorporating Expert Judgment in Economic Forecasts
to a ‘‘green’’ banner in April 1992, indicating that the economy was turning around
and expanding above the long-range growth potential of 3%.
The foregoing provides the context for the forecasting exercises of December
1991 and May 1992 which are described in this paper. As noted, the NBER has
now rendered its judgment concerning the date from which the economic recovery
began. Ensuing events will also disclose the eventual strength of that recovery.
However, as will be illustrated in a later section of this paper, our judgment in
December 1991 was that a meaningful turning point in the current cycle was still a
number of quarters in the future. Our judgment in May 1992, moreover, was that
the strength of the eventual recovery was likely to be quite weak when compared
to previous expansions, owing chiefly to the ‘‘braking’’ influence of major struc-
tural changes then taking place in the domestic and global economies (specifically,
such factors as the de-emphasis of production based on national defense and the
increasing integration of world financial markets).
Our forecasting exercises employed the AHP to address two critical issues germane
to forecasting: the timing and the strength of the expected recovery. The timing
issue required us to incorporate into the forecasting exercise the sequence of global
events of the previous two and a half years. In our view these events had been
forging a restructuring of global resources and institutional arrangements. With
regard to the strength of the recovery, our task was to think through the ways in
which such restructuring acts as a moderating influence on the performance of the
key macroeconomic variables most proximately connected to the U.S. economic
cycle. Our first exercise thus sought to forecast the most likely period for the
turnaround, while the second tried to predict the strength of the ensuing recovery.
As noted, the objective of the first exercise was to forecast the most likely date of a
turnaround. The top level of both exercises consists of the factors representing the
forces or major influences driving the economy. These forces are grouped into two
categories: ‘‘conventional adjustment’’ and ‘‘economic restructuring.’’ Both of
these categories are decomposed into subfactors represented in the second level.
For the timing forecast, the third level consists of time periods in which the
recovery can occur. Figure 13.1 provides a schematic layout used to forecast the
timing of the economic turnaround.
Because conventional adjustment and restructuring are both time dependent
factors, their relative importance had to be established in terms of each of the four
13.4 Application of AHP to the Macroeconomic Forecasting Problem 175
Fig. 13.1 The U.S. holarchy of factors to forecast a turnaround in economic stagnation
contrasting time periods used to compose the forecast time frame. Thus, instead of
establishing a single goal as one does for a conventional hierarchy, we used the
bottom level time periods to compare the two factors at the top. This entailed the
creation of a feedback hierarchy known as a ‘‘holarchy’’ in which the priorities of
the elements at the top level are determined in terms of the elements at the bottom
level, thus creating an interactive loop.
With regard to forecasting the strength of the recovery, we used a standard
format for the hierarchy, beginning with the primary factors of conventional
adjustment and economic restructuring. Their importance for this part of the
exercise was established over a six month period after the turn around. Figure 13.2
is a standard hierarchy, and provides a representation of relevant factors.
Conventional adjustment assumes a status quo with regard to the system of causes
and consequences in the economy. The presumption is that the underlying structure
of the economy is stationary. Forecasting is possible within acceptable ranges of
error. This is achieved by tracing the existing network of stimulus/response patterns
initiated by a perturbation in a fundamental parameter of the economy. In our view,
conventional adjustment can formally be divided into six macroeconomic subfactors
that occupy the second level: consumer spending, investment spending, exports,
indicators of confidence in the economy, fiscal policy, and monetary policy. We
recognize that these subfactors are in some instances interdependent.
Viewed independently, for example, a lowering of interest rates by the Federal
Reserve should induce portfolio rebalancing throughout the economy. In turn,
this should reduce the cost of capital to firms and stimulate investment. Simul-
taneously, it should reduce financial costs to households and increase their dis-
posable incomes. Any resulting increase in disposable income stimulates
consumption and at the margin has a positive impact on employment and GNP.
However, all of this assumes that the linkages of the economy are in place and
are well understood.
176 13 Incorporating Expert Judgment in Economic Forecasts
Recent events in the global economy will exert fundamental changes in the way
the U.S. economy will operate for the next several years and beyond by inducing
an economic restructuring. The Gulf War, the demise of centrally planned econ-
omies in eastern Europe and the former Soviet Union, the integration of western
Europe, the emergence of newly industrialized economies, and the quickening
integration of financial sectors throughout the world are all events which suggest
an economic structure that is not stationary but is undergoing dramatic change.
Prudent recognition of these facts suggests that patience and monitoring of events
are appropriate guidelines for public policy.
With regard to the nature of the current economic restructuring, we specifically
recognized in this exercise the transformation of the financial sector, the reduction
in the defense-based component of the economy, and the changing global com-
petitiveness position of the U.S. economy as additional subfactors in the second
level.
Changes in the domestic economic environment induced by these factors affect
the economy in ways that are not well understood and are too complex to pursue
here. We summarize these effects by estimating the impact of each subfactor on
the expected length of time prior to a turnaround, as well as their impact on the
relative strength of the ensuing expansion.
13.4 Application of AHP to the Macroeconomic Forecasting Problem 177
With respect to the timing of the turnaround, we considered four possible time
periods of adjustment in the third level as a reasonable breakdown of time in
periods long enough to discern change in making the comparisons, but short
enough to consider all possible changes over the two year horizon of the forecast.
These periods were: 3 months, 6 months, 1 year, and 2 or more years, dated from
late December 1991.
With regard to the strength of the expansion, our May 1992 exercise employed
ranges of average real GNP growth. Specifically, we considered the following
possible outcomes: very strong (5.5–6.5%), strong (4.5–5.5%), moderate (3.0–
4.5%), and weak (2.0–3.0%). These ranges represent annualized measures of
percentage change in real gross national product for the first two years of the
recovery. While the ranges are somewhat arbitrary, they generally reflect actual
experiences during various post World War II cyclical expansions.
After decomposing the two problems hierarchically (i.e. the time period of the
expected turnaround and the relative strength of the ensuing recovery), the second
step in the process was to compare the factors as to their relative importance in
affecting each of these questions in terms of their parent factor in the adjacent level
above. Accordingly, comparisons were carried out using the AHP’s nine point scale.
An illustration of the use of this scale to represent judgments proceeded in the
following manner: if conventional adjustment factors were considered to be ‘‘strongly
more important’’ than economic restructuring factors for an economic turnaround to
occur within six months, the number five (5 times) would have been assigned to the
pairwise comparison of conventional adjustment with economic restructuring.
The judgments with regard to the identification of factors as well as the com-
parisons of relative impact and strength of factors were conducted by the authors,
who assumed the role of representative ‘‘experts’’. Obviously, the outcomes are
heavily dependent on the quality of those judgments. As noted, the first exercise
(timing of the turnaround) was conducted during the third week of December,
1991 and refined during first week of January, 1992. The estimation of the strength
of the recovery was conducted during the second week of May, 1992.
Tables 13.2, 13.3, 13.4, 13.5, 13.6 and 13.7 provide the associated matrices of
relative comparisons as well as a limiting and completed ‘‘supermatrix.’’
For example, in Table 13.2, when comparing consumption with investment as a
means of conventional adjustment, consumption is thought to be strongly more
important and a 5 is entered in the first row and third column (1, 3). Its reciprocal
value of 1/5 is entered in the (3, 1) position. On the other hand, when compared with
confidence, consumption is not more important but confidence is strongly more
important and a 1/5 is entered in the (1, 4) position and a 5 in the (4, 1) position. All
other judgments follow this procedure. The vector of weights is derived from the
matrix as the principal eigenvector of the matrix as described in Sect. 13.3.
178 13 Incorporating Expert Judgment in Economic Forecasts
Table 13.2 Matrices for subfactor importance relative to primary factors influencing the timing
of recovery
Panel A: Which subfactor has the greater potential to influence conventional adjustment and how
strongly?
C E I K F M Vector weights
Consumption 1 7 5 1/5 1/2 1/5 0.118
(C)
Exports 1/7 1 1/5 1/5 1/5 1/7 0.029
(E)
Investment 1/5 5 1 1/5 1/3 1/5 0.058
(I)
Confidence 5 5 5 1 5 1 0.334
(K)
Fiscal policy 2 5 3 1/5 1 1/5 0.118
(F)
Monetary policy 5 7 5 1 5 1 0.343
(M)
Panel B: Which subfactor has the greater potential to influence economic restructuring and how
strongly?
FS DP GC Vector weights
Financial sector 1 3 3 0.584
(FS)
Defense posture 1/3 1 3 0.281
(DS)
Global competition 1/3 1/3 1 0.135
(GC)
Note to Tables 13.6 and 13.7: Now we group all the derived vector weights as
columns in the appropriate positions of a matrix of mutual influences known as the
supermatrix. For example, the first vector we derived from the matrix of subfactors
of conventional adjustment is placed in the first column next to the six subfactors and
under conventional adjustment. The factors are listed systematically so that the right
vectors are listed to indicate the impact of the relevant factors on the left on the
factors at the top. The supermatrix, being stochastic (with columns adding to one) is
then raised to limiting powers to capture all the interactions and obtain the steady
state outcome in which all columns within each block of factors are the same. We are
particularly interested in the two identical columns at the bottom left corner of the
matrix of Table 13.7. Either one is given by (0.224, 0.141, 0.201, 0.424).
To obtain the forecast we multiply each value by the midpoint of its corresponding
time interval and add (as one does when evaluating expected values). We have
0:224 1:5 þ 0:151 4:5 þ 0:201 9 þ 0:424 18 ¼ 10:45months
from early Jan. 1, 1992. Note that at times the resulting supermatrix may not be
stochastic which would then require weighting each cluster of factors as it impacts
another cluster at the top.
13.4 Application of AHP to the Macroeconomic Forecasting Problem 179
Table 13.3 Matrices for relative influence of subfactors on periods of adjustment (months)
(conventional adjustment)
3 6 12 24 Weights
Panel A: Relative importance of targeted time periods for consumption to drive a turnaround
3 months 1 1/5 1/7 1/7 0.043
6 months 5 1 1/5 1/5 0.113
12 months 7 5 1 1/3 0.310
24 months 7 5 3 1 0.534
Panel B: Relative importance of targeted time periods for exports to drive a turnaround
3 months 1 1 1/5 1/5 0.083
6 months 1 1 1/5 1/5 0.083
12 months 5 5 1 1 0.417
24 months 5 5 1 1 0.417
Panel C: Relative importance of targeted time investment to drive a turnaround
3 months 1 1 1/5 1/5 0.078
6 months 1 1 1/5 1/5 0.078
12 months 5 5 1 1/3 0.305
24 months 5 5 3 1 0.538
Panel D: Relative importance of targeted time for periods periods for fiscal policy to drive a turnaround
3 months 1 1 1/3 1/5 0.099
6 months 1 1 1/5 1/5 0.087
12 months 3 5 1 1 0.382
24 months 5 5 1 1 0.432
Panel E: Relative importance of targeted time periods for monetary policy to drive a activity to
turnaround
3 months 1 5 7 7 0.605
6 months 1/5 1 5 7 0.262
12 months 1/7 1/5 1 1/5 0.042
24 months 1/7 1/7 5 1 0.091
Panel F: Expected time for a change of confidence indicators of consumer and investor support a
turnaround in the economy
3 months 1 3 5 5 0.517
6 months 1/3 1 5 5 0.305
12 months 1/5 1/5 1 5 0.124
24 months 1/5 1/5 1/5 1 0.054
For each panel, which time period is more likely to indicate a turnaround if the relevant factor is the sole
driving force?
13.4.3 Synthesis/Results
When the judgments were made, the software package known as ‘‘Expert Choice [4],’’
in which the AHP procedure is embedded, was used to perform a synthesis which
produced the following results:
1. A meaningful turnaround in the economy would likely require an additional ten
to eleven months, occurring during the fourth quarter of 1992. This forecast was
derived from weights generated in the first column of the limiting matrix in
180 13 Incorporating Expert Judgment in Economic Forecasts
Table 13.4 Matrices for relative influence of subfactors on periods of adjustment (months)
(Economic Restructuring)
3 6 12 24 Vector weights
Panel A: Most likely length of time for restructuring of financial system to support a turnaround
3 months 1 1/3 1/5 1/7 0.049
6 months 3 1 1/5 1/7 0.085
12 months 5 5 1 1/5 0.236
24 months 7 7 5 1 0.630
Panel B: Most likely time required for defense readjustment to affect a turnaround in economy
3 months 1 1/3 1/5 1/7 0.049
6 months 3 1 1/5 1/7 0.085
12 months 5 5 1 1/5 0.236
24 months 7 7 5 1 0.630
Panel C: Most likely time required for an adjustment to global competition can affect a
turnaround in economy
3 months 1 1 1/3 1/5 0.089
6 months 1 1 1/3 1/5 0.089
12 months 3 3 1 1/5 0.208
24 months 5 5 5 1 0.613
For each panel, which time period is more likely to indicate a turnaround if the relevant factor is
the sole driving force?
Table 13.5 Most likely factor to dominate during a specified time period
CA R Vector weights
Panel A: 3 Months
CA 1 5 0.833
R 1/5 1 0.167
Panel B: 6 Months
CA 1 5 0.833
R 1/5 1 0.167
Panel C: 1 Year
CA 1 1 0.500
R 1 1 0.500
Panel D: 2 Years
CA 1 1/5 0.167
R 5 1 0.833
Which factor is more likely to produce a turnaround during the specified time period?
CA conventional adjustment
R restructuring
Table 13.7, coupled with the mid-points of the alternate time periods to provide
unbiased estimates:
2. At an annual percentage change in real gross national product of about 3.6%,
the recovery would be ‘‘moderate’’ (using our range definition). Tables 13.8,
13.9, 13.10, 13.11 provide the relevant comparison matrices. Specifically,
Table 13.11 documents the judgments regarding the strength of the expansion.
13.4
2 years 0.0 0.0 0.534 0.417 0.539 0.054 0.432 0.091 0.630 0.630 0.613 0.0 0.0 0.0 0.0
181
182
Investment 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.028 0.028 0.028 0.028
Confidence 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.162 0.162 0.162 0.162
Fiscal policy 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.057 0.057 0.057 0.057
Monetary policy 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.166 0.166 0.166 0.166
Financial sector 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.301 0.301 0.301 0.301
Defense posture 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.145 0.145 0.145 0.145
Global competition 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.070 0.070 0.070 0.070
3 months 0.224 0.224 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
6 months 0.151 0.151 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1 year 0.201 0.201 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
2 years 0.424 0.424 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Incorporating Expert Judgment in Economic Forecasts
13.4 Application of AHP to the Macroeconomic Forecasting Problem 183
Table 13.8 Matrices for primary and subfactors for strength of recovery
Panel A: Which primary factor will be more influential in determining the strength of the
recovery?
CA R Vector weights
Conventional adjustment (CA) 1 1/5 0.167
Restructuring (R) 5 1 0.833
Panel B: Which subfactor is more important in influencing conventional adjustment?
C E I K F M Vector weights
Consumption (C) 1 7 3 1 7 3 0.3117
Exports (E) 1/7 1 1/5 1/5 1 1/5 0.037
Investment (I) 1/3 5 1 1/3 1/3 1/5 0.099
Confidence (K) 1 5 3 1 7 3 0.305
Fiscal policy (F) 1/7 1 3 1/7 1 1/7 0.035
Monetary policy (M) 1/3 7 5 1/3 7 1 0.207
Panel C: Which subfactor is more important in influencing economic restructuring?a
FS DP GC Vector weights
Financial sector (FS) 1 1/5 1/3 0.105
Defense posture (DS) 5 1 3 0.637
Global competition (GC) 3 1/3 1 0.258
a
CI = 0.037
Note to Tables 13.8, 13.9, 13.10, 13.11: The next example is a simple hier-
archy as shown in Fig. 13.2 in which the derived weights in each level starting at
the top are weighted by the weight of the corresponding factor in the level above
used to compare the elements. The results are then added for each element to
obtain its overall weight. The process is continued to the bottom level of the
hierarchy.
13.5 Conclusion
This chapter has demonstrated how the Analytic Hierarchy Process can serve as an
additional tool for macroeconomic forecasts. We have used the highly interesting
and relevant case of the U.S. economy during its current economic cycle (in which
structural change has been particularly important) as the specific context for our
analysis. As noted earlier, this approach could easily be adopted for use in forecasts
based initially on formal macroeconometric models (e.g. to make judgments on
shifts in intercepts and changes in the value of exogenous variables).
With regard to our forecasts, in addition to presenting the somewhat con-
trarian view that a meaningful turnaround in the present economic cycle was then
still some months in the future, we concluded that the next recovery would be
substantially less strong than those of the past three decades. We viewed this as
184 13 Incorporating Expert Judgment in Economic Forecasts
Table 13.9 Matrices for relative influence of subfactors on strength of recovery (conventional
adjustment)
V S M W Vector weights
Panel A: Relative likelihood of the strength of recovery if consumption drives the expansiona
Very strong (V) 1 1 5 7 0.423
Strong (S) 1 1 5 7 0.423
Moderate (M) 1/5 1/5 1 3 0.104
Weak (W) 1/7 1/7 1/3 1 0.051
Panel B: Relative likelihood of the strength of recovery if exports drives the expansionb
Very strong (V) 1 1 1/3 1/5 0.095
Strong (S) 1 1 1/3 1/5 0.095
Moderate (M) 3 3 1 1/3 0.249
Weak (W) 5 5 3 1 0.560
Panel C: Relative likelihood of the strength of recovery if investment drives the expansionc
Very strong (V) 1 1 1/3 2 0.182
Strong (S) 1 1 1/3 2 0.182
Moderate (M) 3 3 1 6 0.545
Weak (W) 1/2 1/2 1/6 1 0.091
Panel D: Relative likelihood of the strength of recovery if confidence drives the expansiond
Very strong (V) 1 1 3 5 0.376
Strong (S) 1 1 3 5 0.376
Moderate (M) 1/3 1/3 1 7 0.193
Weak (W) 1/5 1/5 1/7 1 0.054
Panel E: Relative likelihood of the strength of recovery if fiscal policy drives the expansione
Very strong (V) 1 1 1/5 1 0.125
Strong (S) 1 1 1/5 1 0.125
Moderate (M) 5 5 1 5 0.625
Weak (W) 1 1 1/5 1 0.125
Panel F: Relative likelihood of the strength of recovery if monetary policy drives the expansionf
Very strong (V) 1 1 1/5 1/3 0.084
Strong (S) 1 1 1/5 1/3 0.084
Moderate (M) 5 5 1 7 0.649
Weak (W) 3 3 1/7 1 0.183
For each panel, which intensity is more likely to obtain if the designated factor drives the
recovery?
a
C.I. = 0.028
b
C.I. = 0.016
c
C.I. = 0.0
d
C.I. = 0.101
e
C.I. = 0.0
f
C.I. = 0.101
Table 13.10 Matrices for relative influence of subfactors on strength of recovery (restructuring)
V S M W Vector weights
Panel A: Relative likelihood of the strength of recovery if financial sector drives the expansiona
Very strong (V) 1 1 1/3 1/5 0.095
Strong (S) 1 1 1/3 1/5 0.095
Moderate (M) 3 3 1 1/3 0.249
Weak (W) 5 5 3 1 0.560
Panel B: Relative likelihood of the strength of recovery if defense posture drives the expansionb
Very strong (V) 1 1/3 1/5 1/7 0.055
Strong (S) 3 1 1/3 1/5 0.118
Moderate (M) 5 3 1 1/3 0.262
Weak (W) 7 5 3 1 0.565
Panel C: Relative likelihood of the strength of recovery if global competition drives the
expansionc
Very strong (V) 1 1 1/3 1/5 0.101
Strong (S) 1 1 1/3 1/5 0.101
Moderate (M) 3 3 1 1 0.348
Weak (W) 5 5 1 1 0.449
For each panel, which intensity is more likely to obtain if the designated factor drives the
recovery?
a
CI = 0.016
b
CI = 0.044
c
CI = 0.012
Bibliography
1. Adams FG (1986) The business forecasting revolution. Oxford University Press, Oxford
2. Council of Economic Advisors (1993) Economic Report of the President
3. Fair RC (1984) Specification, estimation, and analysis of Macroeconometric models. Harvard
University Press, Cambridge
4. Forman E, Saaty TL Expert Choice Software, produced by Expert Choice, Inc., 4922
Ellsworth Avenue, Pittsburgh, PA
5. Hall TE (1990) Business cycles—The nature and causes of economic fluctuation. Praeger,
New York
6. Hershey RD Jr (1992) June 18, Good Riddance to Recession? New York Times, 1992
7. Hershey RD Jr, (1992) December 23, This Just In: Recession Ended 21 Months Ago, New
York Times, 1992
186 13 Incorporating Expert Judgment in Economic Forecasts
8. Miller GA (1956) The magical number seven plus or minus two: some limits on our capacity
of processing information. Psychol Rev 63:81–97
9. Saaty TL, Vargas LG (1991) Prediction, projection and forecasting. Kluwer Academic,
Boston
10. Saaty TL (1990) Multicriteria decision making: the analytic hierarchy process. RWS
Publications, Pittsburgh
11. Saaty TL, Alexander JM (1989) Conflict resolution: the analytic hierarchy approach. Praeger,
New York
12. Saaty TL, Kearns KP (1985) Analytical planning: the organization of systems. Pergamon
Press, New York
13. Sorkin AL (1988) Monetary and fiscal policy and business cycles in the modern Era.
Lexington Books, Lexington
14. U.S. Department of Commerce, Survey of Current Business, various issues
Chapter 14
A New Macroeconomic Forecasting
and Policy Evaluation Method
14.1 Introduction
The economy is often faced with a turn that is not to our liking, and we sometimes
think it ought to be controllable by macroeconomic policy. While the spectrum of
policies ranges from Keynesian ‘‘fine tuning’’ to using monetarist ‘‘rules,’’ every
action (or inaction) of the government is a policy. Because of this it is important
that government policy makers be guided by appropriate empirical models.
Unfortunately, there is considerable variation in the large numbers of econometric
models that have been developed thus far. Moreover, there are persistent problems,
both technical and theoretical, with these models.
We will show that the Analytic Hierarchy Process (AHP) is an effective method
for forecasting the end effects of a given policy or set of policies, and for deter-
mining the resulting impact on important variables such as unemployment and
inflation. The forecasts could be made using the judgments of leading economists,
congressmen, and personnel from major federal agencies such as OMB, CEA,
commerce, the Federal Reserve and the Treasury. One side benefit would be a
clearer understanding and appreciation of the problem under consideration as
viewed from these different perspectives. The analysis in this chapter reflects the
economic climate of the early 1980s.
There are problems with currently existing econometric models. Some problems
are purely technical in nature and arise from the statistical approach. For example,
there is a need to determine the proper estimation technique for equations with
lagged dependent variables and serially correlated error terms when these equa-
tions are embedded in a simultaneous system of equations. There has been steady
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 187
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_14, Springer Science+Business Media New York 2012
188 14 A New Macroeconomic Forecasting and Policy Evaluation Method
progress in the development of these statistical models but their formalism does
not yield satisfactory solutions.
Other problems with the current crop of econometric models are more serious.
For instance, the presence of so many models, all differing significantly in their
structure, raises the question about how to specify the type of model and the
parameters. There is a yet more serious question common to virtually all econo-
metric models that use time series data to estimate historical correlations between
variables. The estimated coefficients that represent peoples’, firms’, and govern-
ments’ behavior are typically presumed to be ‘‘structural’’ in nature-that is, they
are assumed to be invariant with respect to changes in the economic environment
and, in particular, to changes brought about by policy actions. But, instead of being
constant, these coefficients are unstable and may account for ‘‘sudden, unpre-
dictable shifts in behavioral relationships…’’ [2]. Economists using these models
have responded by developing statistical techniques that allow the estimated
parameters to drift at random rather than forcing them to be constant [4, 6]. This
has resulted in forecasts of somewhat increased accuracy [12].
However, allowing for random parameters in the estimation process is at best a
statistical device for coping with a more basic problem. A fundamental observa-
tion is that on the part of consumers only the parameters of utility functions are
truly invariant with respect to changes in policy, while for firms the same is true
only for the parameters of their production functions. All ‘‘Behavioral’’ coeffi-
cients in an individual‘s demand and supply functions depend, in some fashion, on
more primitive parameters in his utility function and his perceived budget con-
straint. After all, demand functions are the result of the individual maximizing his
utility subject to the restraint given by the perceived economic environment as
embodied in his budget constraint. Any change in government policy alters the
economic environment and leads to a change in the structure of the demand
function as the individual responds to the changed incentives. The same general
conclusions are true for firms’ (and individuals’) supply functions. As a result, the
assumptions made in most econometric models that the estimated coefficients of
the demand and supply functions are stable with respect to changes in policy are
hard to defend. Sargent [9] discusses in detail the econometric consequences of
these observations (which originated with Lucas [5]) for standard econometric
models.
These considerations have two immediate implications. The first is that the
‘‘sudden, unpredictable shifts…in the coefficients‘‘ are no longer a puzzle. The
instability is a manifestation of the fact that a forecasting model‘s coefficients are
not truly structural in nature and, because of this, shifts and changes are to be
expected. The second, and more serious, weakness is that standard econometric
models are particularly unable to forecast results of alternative policies because as
different policies are instituted the parameters of the estimated demand and supply
functions change due to people and firms altering their behavior. In virtually all
econometric models, this effect is simply ignored because of technical difficulties.
These defects of existing models have led to widely differing reactions among
economists. Equilibrium theorists tend to regard the objections sympathetically
14.2 A Few Words about Existing Econometric Models 189
We will consider the problem of finding the best macroeconomic policy package
among a given set by considering the effect that each policy package has on
national ‘‘welfare.’’ The effect of each package on national welfare is, in turn,
decomposed into the effects of the package on key criteria such as inflation,
unemployment, and growth. In terms of the AHP, this problem formulation leads
to a three level hierarchy. The goal of best policy for the national welfare is in the
first level. The criteria occupy the second level and the policy packages occupy the
third level. The hierarchy, corresponding to this problem from which the relative
effectiveness of the different policies on the third level are determined, is illus-
trated later.
In the second stage we incorporate both the uncertainty concerning the exogenous
forces operating on the economy and the possibility that the effectiveness of the
various policies may change over time. Uncontrollable forces that affect the
economy are included in the hierarchy in another level that involves what we call
‘‘exogenous scenarios,’’ or SE’s. These SE’s consist of elements beyond the
control of the policymaker that are deemed to exert an important effect upon the
economy. Examples of such elements are drastic changes in the real price of
energy, or wars involving vital interests of the nation. Since these scenarios are by
their nature uncertain, a number of potential scenarios are constructed. Moreover,
time is intimately involved whenever a scenario is constructed, and the expected
SE may change with the passing of time.
190 14 A New Macroeconomic Forecasting and Policy Evaluation Method
the weighted vectors are summed to arrive at the overall expected relative con-
tribution of the policies toward achieving the nation’s welfare. Since each of these
summed vectors applies to time periods somewhere in the future, they are dis-
counted to the present to arrive at the overall effectiveness of the policies. One
selects the policy with the largest priority.
After the policies have been qualitatively ranked, the next step is to obtain the
quantitative forecasts. We will illustrate our forecasting procedure by using only
the first stage, short-run technique.
The forecasting procedure introduces a second hierarchy, which extends the
first hierarchy further down, and adds a new element ‘‘non-controllable actors.’’
Non-controllable actors are organizations, groups of individuals, or institutions
who wield substantial influence over certain sectors or variables but who are not
under the direct control of the policy maker. The non-controllable actors will
occupy the second level in this hierarchy, immediately beneath each criterion. The
level under the non-controllable actors will contain the range of actions they can
take to influence the criteria. Immediately beneath this will be the level containing
the quantitative effect each possible action could have on the criteria.
An example of this type of hierarchy is given in Fig. 14.2 for the criterion of
inflation. The second row contains suggested non-controllable actors who exert
192 14 A New Macroeconomic Forecasting and Policy Evaluation Method
direct influence on the inflation rate. It would be possible to include other groups
who can influence the inflation rates but the three given in the figure will suffice for
the purposes of this illustration. The first step would be to construct a comparison
matrix between the three non-controllable actors and to calculate the relative
influence each exerts upon the inflation rate.
The third level in Fig. 14.2 illustrates a range of possible actions consumers
could take that would affect the inflation rate. There would be other analogous
entries in this level for the other non-controllable actors. We have omitted them to
simplify things. These five actions are grouped in a comparison matrix and ranked
according to the question: ‘‘Given that whatever specific policy we are quantifying
has been adopted, what is the likelihood that one of the actions (e.g., increase
demand with moderate intensity, ‘‘moderately’’) will occur relative to another
action?’’ The weights obtained from this matrix will represent the relative likeli-
hoods that consumers will change their demand for goods by the given intensity.
The last level in the hierar-chy lists the range of possible inflation rates. Again,
each entry in the ‘‘possible actions’’ level of the hierarchy will have associated
with it a range of possible inflation rates. Again, for simplicity, we do not show all
the detail. This level is prioritized in a comparison matrix by answering the
question: ‘‘Given that the specific possible action has occurred, (e.g., no change in
demand) what is the likelihood that inflation will assume a certain value as
compared to another?’’ The weights obtained from this exercise will be the pri-
orities that the inflation rate will be equal to the given amounts.
The next step is to obtain the quantitative forecast. This is done by first
weighting the possible inflation rates by the priority of their occurrence for each
possible action. These expected inflation rates are then weighted by the likelihood
that the possible action to which they pertain takes place. For example, the
expected inflation rate, given that there is no change in demand by consumers,
might be 7%. This figure is, in turn, weighted by the priority that there is no
change in demand by consumers. When this has been done for each of the possible
actions by consumers, the (five) weighted figures are summed to arrive at the
expected value of the ‘‘consumer’s contribution’’ to the inflation rate. This pro-
cedure is carried out for all other non-controllable actors being considered. The
final step is to weight each non-controllable actor’s expected contribution by the
relative importance of the actor to arrive at the overall expected inflation rate. This
procedure is implemented for all the criteria to be forecast for all the specific
policies being considered.
There are several useful points to make about this approach. First, in other cases in
which forecasting has actually been carried out, the results have proven to be accurate
by other measures (See Zahedi [11] for forecasting applications). In addition, when
forecasting, each comparison matrix is based on the specific policy considered. The
question used to rank the possible actions refers directly to the policy package for
which the forecast is being made. It is at this stage that we allow the non-controllable
actors to react differently according to whatever policy we are considering. Thus we
take into account the criticism of Lucas and others that the behavior of economic
agents depends strongly upon the policies that are in force.
14.2 A Few Words about Existing Econometric Models 193
Finally, the use of the AHP, combined with judgment based comparisons
eliminates the complaint levied against econometric models that the results of
policy exercises are predetermined by the model‘s basic structure. The AHP is
sufficiently general so that it is possible to include a variety of non-controllable
actors in an effort to capture effects neglected by specific models and modelers.
Economists belonging to the monetarist school tend to dismiss the effect that
organized labor has on the inflation rate. If it were a monetarist economist using
the AHP to forecast the inflation rate, it would still be possible to include labor
unions in the series of non-controllable actors—the monetarist would simply give
it a small weight.
hierarchy. The criteria occupy the second level and the policy packages, to be
evaluated for how they affect each criterion, are in the bottom level.
In our example, we took the vantage point of a ‘‘representative’’ person in
society to provide judgments for the pairwise comparisons of the criteria with
respect to their importance to national welfare. It should be emphasized once again
that any attempt through any model to develop economic policy must specify
implicitly or explicitly an objective function for that policy. Either way such an
objective function will rely on value judgments. Thus, establishing the relative
importance of the criteria in level one for a ‘‘representative’’ person in society
amounts to specifying explicitly the objective function and the value judgments
leading to it. The result is the vector of relative weights, or the relative importance
of the criteria to national welfare, given in Table 14.l. Note that a reduction in the
inflation rate is deemed the most important action to improve national welfare. A
reduction in unemployment, while not as important as a reduction in the inflation
rate, is still more important than the remaining criteria. The other criteria have very
low priorities in comparison with inflation and unemployment. The values in this
vector of relative weights can be interpreted either as the importance of one
criterion over another, e.g., inflation is nine times (0.45/0.05) as important as
foreign relations, or as the relative attention that should be paid a particular cri-
terion (inflation = 45%) in attempting to improve national welfare.
The next level of the hierarchy involves specifying policy packages and fore-
casting their effect upon each of the five criteria. Various new and traditional
macropolicies were selected, 20 in all. For simplicity, the policies were clustered
into three sets ‘‘A’’, ‘‘B’’, and ‘‘C’’, corresponding to what is generally regarded as
‘‘conservative,’’ ‘‘moderate’’ and ‘‘liberal’’ respectively. The policy packages were
evaluated with respect to the economic criteria (inflation, unemployment and
growth), three times–from the vantage point of a monetarist, from that of a
Keynesian and, finally, from a ‘‘supply-side’’ viewpoint. For the criteria of
domestic relations and foreign relations we evaluated the impact of the policy
packages from the point of view of a ‘‘representative’’ individual again; it was not
necessary to distinguish among monetarist, Keynesian, and supply-sider opinions.
We then selected the optimal policy from each set (A, B, and C) for each economic
viewpoint, combined the optimal policies into a single set and selected the best
overall policy for each school.
The policies were evaluated only for their presumed short-run effect on the five
criteria. And we took into consideration the constraint that the government’s
budget places on the use of various policy tools. For example, if the government
14.3 Application of the AHP to Macroeconomic Policy 195
lowers tax rates by 10% while leaving its total spending unchanged, there would
be a potential shortfall of revenue. If this were the case, either bonds and/or money
would have to be issued to finance the tax cut and to reconcile expenditures with
revenues. Thus, in the following development, we specified complete policies
(e.g., ‘‘lower tax rates by 10 percent and issue money to cover any shortfall’’). The
notation and symbols to describe the policies are given in Table 14.2.
In Table 14.3, the evaluation of cluster B policies with respect to inflation from
the monetarist viewpoint, we give an example of how we evaluated cluster policies
through pairwise comparisons with respect to a criterion. Similar comparisons
were carried out for every economic group and every cluster. The judgments were
given in response to the following type of question: ‘‘From a monetarist viewpoint,
how much more effective in reducing the inflation rate will be a policy of
decreasing tax rates by 10% while balancing the budget using money and bonds
196 14 A New Macroeconomic Forecasting and Policy Evaluation Method
Table 14.4 Determination of the overall effectiveness of the policy (G, Tx, M; 10%)
(Monetarist viewpoint)
Objective A Weight of the Weight of the policy with Contribution to the overall
criterion B respect to the criterion C criterion D = B 9 C
Inflation 0.450 0.120 0.05
Unemployment 0.300 0.103 0.03
Growth 0.110 0.306 0.03
Domestic 0.090 0.240 0.02
stability
Foreign 0.050 0.099 0.01
relations
Overall effectiveness 0.14
(i.e., Tx;, B, M: 10%) than will be a policy of cutting tax rates by five and issuing
money and bonds to make up any revenue shortfall (i.e., Tx;, B, M: 5%)?’’ In this
case, where the policies are ranked from a monetarist perspective, we judged the
second policy to be between strongly and very strongly more effective (i.e., 6) than
the first policy. Since the comparison is, by convention, between the first policy
relative to the second policy, we placed the ranking (1/6) in the matrix. The
remaining comparisons in this matrix were entered in a similar fashion. The
extreme right column gives the vector of relative weights for these four policies to
achieve a reduction in inflation.
Once the policy matrices are prioritized with respect to each of the five criteria,
the overall efficacy of each policy is determined by multiplying the effectiveness
of each policy with respect to a given criterion by the weight of the criterion
toward affecting national welfare, and then adding over all the criteria.
An example of this procedure is given in Table 14.4. In this table we calculated
the overall effectiveness of policy (G, Tx, M; 10%) when judged from a mone-
tarist viewpoint. Column 2 gives the weights of the criteria we developed in
Table 14.1. Column 3 reports how effective this particular policy is in satisfying
the various criteria. The fourth column weights the effectiveness of the policy.
These weighted contributions are summed and the overall effectiveness of the
policy is given in the bottom row of the table.
Now we have an initial sorting of the 20 policies. The next step is to select the
most effective policies from each cluster and then rank them directly against each
other to find the overall best policy (Tables 14.5–14.7). The technique of clustering
the policies into three groups, calculating the rankings within the groups and then
calculating the rankings between the most effective policies within the groups
might appear to involve more work than directly comparing all 20 potential pol-
icies simultaneously. This is, however, not the case as a simple calculation will
show. Each comparison matrix involves n(n - 1)/2 judgments for the pairwise
comparisons where n is the number of elements being compared. Our clustering
method of evaluating policies required a total of 360 comparisons for each
viewpoint. Directly comparing all the policies would have required 500
14.3 Application of the AHP to Macroeconomic Policy 197
14.4 Conclusion
Monetarists consider a policy to reduce the growth rate of the money supply by 6%
matched by a decrease in government spending to be optimal. Keynesians find a
policy of increasing government spending by 10% while financing the deficit by
198 14 A New Macroeconomic Forecasting and Policy Evaluation Method
issuing bonds and money to be the best policy. Supply-side economists favor a
policy to reduce government spending and tax rates by 10% and would allow
change in the money supply to be a balancing factor in the government’s budget
constraint.
It is interesting to see what we can observe about this analysis from the
hindsight of our early 1987 perspective. Probably the best we can say is that a
mixed set of policies prevailed. There was a supply-sider in the White House
(Ronald Reagan) and something of a monetarist (Paul Volcker) holding the reins at
the Fed. Taxes were decreased; inflation has moderated significantly; growth in the
economy has been modest but steady since late 1982. However, it has not been
possible to restrain government spending, thought by both monetarists and supply-
siders in 1981 to be a very desirable end to achieve. This was the result primarily
of an impasse between Congress and the Administration as to where to make the
cuts—defense or social programs. (The recently enacted Gramm-Rudman legis-
lation seeks to address this problem by annual across-the-board cuts if agreement
is not forthcoming). There has accordingly been a tremendous rise in the budget
deficit and in the trade deficit; unemployment, a goal though still high relative to
the earlier post-war period, has been brought under control; interest rates are down
significantly; the stock market has more than doubled from a Dow Jones Index of
less than 1000 to over 2000. Whether this signifies a healthy economy remains to
be seen and is probably dependent on one’s point of view. The average U.S. citizen
may consider himself to be better off than he was in 1981, but to those outside the
U.S., we are living on borrowed money—to finance our trade and budget deficits.
14.4
a
Table 14.7 Overall effectiveness of policies in cluster C
Policies (%) Overall effectiveness
Monetarist viewpoint Keynesian viewpoint Supply side viewpoint
Conclusion
In any event we have now shown how the AHP can be applied to the problem of
selecting an optimal macroeconomic policy and estimating its impact.
This approach to forecasting has some advantages over more conventional
methods. For instance, it makes it possible to incorporate ideas or theories into
forecasts that are difficult to include in quantitative econometric models. In
addition, since the AHP procedure is essentially ‘‘structure- free,’’ it eliminates the
criticism that the results obtained from its use are predetermined by its a priori
14.4 Conclusion 201
specification. Our brief analysis shows that the optimal choices of the various
schools can be predicted by this process. The AHP also shows how strongly each
ranks alternative policies relative to their primary preferences.
Bibliography
1. Anderson LC, Jordan JL (1968) Monetary and fiscal action: a test of their relative importance
in economic stabilization. Federal Reserve Bank of St. Louis, Review, Nov 1968
2. Cuthbertson K (1979) Macroeconomic policy: the new Cambridge, Keynesian and monetarist
controversies. Wiley, New York
3. Harker PT (1987) Alternative modes of questioning in the analytic hierarchy process. Math
Model 9(3–5):353–360
4. Hildreth C, Houck J (1968) Some estimators for a linear model with random coefficients.
J Am Stat Assoc 63:584–595
5. Lucas RE (1976) Econometric policy evaluation: a critique, in the Phillips curve and labor
markets. In: Brunner K, Meltzer AH (eds) The Carnegie-Rochester con-ferences on public
policy, vol 1. North-Holland, Amsterdam, p 20
6. Maddala GS (1977) Econometrics. McGraw Hill International, New York
7. Saaty TL (1980) The analytic hierarchy process. McGraw Hill International, New York
8. Saaty TL (1987) A new macroeconomic forecasting and policy evaluation method using the
analytic hierarchy process. Math Model 9(3–5):219–232
9. Sargent TJ (1981) Interpreting economic time series. J Political Economy 89(2):213–248
10. Sims CA (1980) Macroeconomics and reality. Econometrica 48(1):1–48
11. Zahedi F (1986) The analytic hierarchy process—A survey of the method and its
applications. Interfaces 16(4):96–108
12. Zarnowitz V (1978) On the accuracy and properties of recent macroeconomic forecasts. Am
Econ Rev 68(2):313–321
Chapter 15
A New Approach to the Middle East
Conflict: The Analytic Hierarchy Process
15.1 Introduction
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 203
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_15, Springer Science+Business Media New York 2012
204 15 A New Approach to the Middle East Conflict
Some of the world’s best negotiators, diplomats and able leaders have grappled
with the resolution of this conflict. However, despite their best efforts, the current
condition continues to torment all the parties. Since the inception of the Analytic
Hierarchy Process and its generalization to dependence and feedback, the Analytic
Network Process (ANP), authors have conducted numerous case studies (e.g. [1–3])
that suggest the method as an alternative approach to conflict resolution that will lay
bare the structure of the problem and allow reasoned judgment to prevail.
Nonetheless, when one deals with conflict, especially conflict of a prolonged
duration, reason rarely prevails. In fact, with respect to the conflict between the
Palestinians and the Israelis, positions have become entrenched and each party
seeks not only to satisfy its own needs but also does not mind increasing the costs
of concessions made by the other party. This type of conflict is defined as
retributive (Saaty 1986) because of its prolonged negative emotional content.
Retributive responses differ from the usual cooperative conflicts in which the
parties work for a win–win outcome, by their partly malevolent intentions,
whereby the parties do not care about the losses of the other side.
In most long-lasting conflicts, each party’s grievances increase while the
concessions they are willing to make decline in number, quality, and perceived
value. Both parties lose sight of what they are willing to settle for, generally
exaggerate their own needs, and minimize the needs of the other side over time.
The concessions worth trading versus the concessions the other party is willing to
trade become more indefinite and less concise. But, it is precisely the matter of
trading that needs to be made more concrete and of higher priority for both sides, if
a meaningful resolution is to be found.
Without a formal way of trading off the concessions and packages of conces-
sions, both sides are likely to suspect that they are getting the short end of the
bargain. After the parties have agreed to a trade, very specific binding language
about the terms of the agreement, clear implementation policies and outside
guarantors are needed. The worth of the concessions traded, as perceived by both
the giver and receiver, need to be accurately determined and recorded. All of this
requires going beyond verbal descriptions of the concessions to more broadly
include their economic, social, geographic, humanitarian and historical worth. It is
critical that all of this needs to be translated into priorities derived in terms of the
different values and beliefs of the parties. Priorities are universal and include the
diversity of measures in terms of which economic, social and other values are
measured. The Analytic Hierarchy Process (AHP) provides a way to perform such
an assessment with the participation of negotiators for the parties. It is a positive
approach that makes it possible to reason and express feelings and judgments with
numerical intensities to derive priorities.
With the assistance of the panel of Israeli participants and Palestinian partici-
pants, AHP has now been applied for the first time with the input of representatives
of both sides who were knowledgeable and informed about the issues associated
with the Palestinian–Israeli conflict. They obviously did not represent the full
spectrum of political ideas and notions. The process makes it clear that moderation
in different degrees by both sides is essential to arrive at acceptable agreements on
15.1 Introduction 205
concessions proposed and agreed upon by both sides. AHP makes it possible to
evaluate moderate and extreme viewpoints and determine their effect on the
trading of concessions. The results obtained encourage us to advocate its use in
this negotiation process.
We need to begin by emphasizing that the outcome of our effort is the begin-
ning of an elaborate undertaking to produce a viable solution to the Israeli–
Palestinian conflict. It is simply a novel framework for dialogue. A differentiation
from other approaches is its potential to minimize the influence on the outcome of
much of the intense emotions that have usually accompanied such discussions. The
framework forces the negotiators to approach the issues using a quantitatively
oriented set of judgments to compare and tradeoff various issues, benefits, costs
and concessions in a way in which each individual item is separated from the
influences of other passionately charged items. We acknowledge that in an emo-
tionally charged conflict such as this, there will inevitably remain a residual
emotionality and feelings that cannot be ignored and inevitably affect the judg-
ments. This does not affect the viability of the process, because the numerical
representation of the judgments allows for such variability up to a limit that can be
measured. It essentially allows one to decompose the problem into smaller
components that can be dealt with more easily. While judgments may vary
according to the perceived power of the parties, the essential nature of the process
is not compromised, unless participants are influenced to change their judgments.
One might ask: Why is it that so many distinguished politicians and negotiators
have failed to reach consensus after 60 years of trying? Here are some possible
reasons:
1. They had no way to measure the importance and value of intangible factors
which can dominate the process.
2. They had no overall unifying structure to organize and prioritize issues and
concessions.
3. They had no mechanism to trade off concessions by measuring their worth.
4. They had no way to capture each party’s perception of the other side’s benefits
and costs.
5. They had no way to provide confidence for the other party that the opposing
party is not gaining more than they are.
6. They had no way to avoid the effect of intense emotions and innuendoes which
negatively affect the negotiation process.
7. They had no way to test the sensitivity and stability of the solution to changes in
their judgments with respect to the importance of the factors that determined
the best outcome.
206 15 A New Approach to the Middle East Conflict
The AHP is about breaking a problem down and then aggregating the solutions of
all the sub-problems into a conclusion. It facilitates decision making by organizing
perceptions, feelings, judgments, and memories into a framework that exhibits the
forces that influence a decision. In the simple and most common case, the forces
are arranged from the more general and less controllable to the more specific and
controllable. The AHP is based on the innate human ability to make sound
judgments about small problems and also about large problems when a structure
like a hierarchy can be built to represent the influences involved. It has been
applied in a variety of decisions and planning projects in nearly 40 countries.
Briefly, we see decision making as a process that involves the following steps:
(1) Structure a problem with a model that shows the problem’s key elements and
their relationships
(2) Elicit judgments that reflect knowledge, feelings, or emotions of the primary
parties, as well as all other parties that have influence on the outcome
(3) Represent those judgments with meaningful numbers
(4) Use these numbers to calculate the priorities of the elements of the hierarchy
(5) Synthesize these results to determine an overall outcome
(6) Analyze sensitivity to changes in judgment
The retributive conflict resolution approach presented here takes into consid-
eration the benefits to A from concessions by B and the costs to A of the return
concessions A makes, as well as A’s perception of the benefits to B from the
concessions A makes, and also A’s perception of the costs to B of the concessions
B makes. A similar consideration is made for B. Findings from this exercise
suggest that the development of ‘‘bundles’’ of concessions may minimize the
difference in ratios of gains and losses between the two parties that a negotiator
can use as a tool to move the resolution process forward.
The expressed objectives of the study were:
• To identify the issues, major and minor and to examine the relative significance
or priority of the issues currently inhibiting solution of the Israeli–Palestinian
conflict
• To share knowledge and insights about the current Israeli–Palestinian situation
from differing points of view
• To construct a comprehensive model of the situation
• To explore the benefits and costs of alternative courses of action
15.3 The Process 207
The traditional approach involving diplomacy and face to face negotiations has
led to an inconclusive outcome, partially attributable to attitudes colored by strong
emotions on both sides. Our approach attempts to address the impact of negative
attitudes by focusing the participants on making judgments that measure the
intensity of their perceptions about the influences that each of the issues brings to
bear upon the final outcome.
In this study we consider each party’s list of issues, which if addressed by the
other party by making concessions, would provide sufficient benefit to that side
towards meeting their goal. They, in turn, would be willing to make concessions to
the other side to balance those concessions with an equivalent tradeoff. We refer to
these issues as criteria. The process consists of taking a set of concessions from
one side and measuring them against these criteria in terms of actual or perceived
benefits to the other side. Actual benefits (or costs) are defined as judgments by
one party about the relative importance of the concessions they receive (or give).
Perceived benefits (or costs) are defined as putting oneself in the shoes of the other
side to estimate the benefits (or costs), even though that side may have a totally
different opinion about what the concessions received or offered are worth.
The remainder of this work is structured as follows: In the next section we
define the problem in general terms. Subsequently we outline the structure of the
decision in the form of multiple hierarchies. This effect is evaluated by the parties
according to their value systems, both actual and perceived. The outcomes of this
analysis are priorities used to assess ratios of gains and losses by both sides that
make it possible to determine those concessions for which each party’s gains
exceed its losses and these gains to losses are not unacceptably large for either
party in comparison with the other party. Then we examine and identify ratios that
are nearly equal for the two sides from the concessions made, and pose questions
about the viability of such bundles of concessions that are traded off. Finally we
suggest a way for moving the process to the next level through better definition of
the issues and concessions as well as through recognition of potential imple-
mentation policies and other relevant changes.
In the opening day of the 3 day meeting the panel brainstormed the issues and
structured the problem, defined the parties at interest and developed a series of
concessions that each party might offer to the other.
The process was not without conflict and negotiation of its own. At times, the
panel made judgments without agreement on exact definitions. There was nearly
always unanimous agreement on the nature of the conflict, with much debate about
the underlying concerns. These concerns differed according to which constituent
group was putting them forward. For example, among the Palestinian key con-
stituents are Palestinian refugees, Hamas followers, Fatah followers, Palestinians
who still live in Israel and Diaspora Palestinians. Among the Israeli constituents
208 15 A New Approach to the Middle East Conflict
are the ultra right orthodox community, Israelis living in settlements in the West
Bank, those associated with the Likud movement, those associated with the Labor
Party, and those more actively seeking peace as a primary objective, without
dwelling on the details of the difficulties to achieve it.
Since the beginning of the conflict, different constituents have proposed many
different approaches. These approaches inevitably influenced the panel’s perception
of the concessions to be made by either side. In fact, one participant suggested that
it would be difficult ‘‘to think outside the box.’’ He thought that the group was so
influenced by previous thinking that they would have difficulty in conceptualizing
‘creative’ alternatives that had not been proposed previously.
The panel defined the goal as an attempt to understand what forces and influ-
ences or combinations thereof would tend toward a consensus peace accord for the
conflict between Israel and the Palestinians. To accomplish this goal, the panel of
nine individuals was assembled to represent a cross section of thinking on both
sides. Its members had present or prior experience in academia, government and in
business. However, it was recognized that the panel did not represent a complete
cross-sample of opinions. The sample of panel participants was not sufficiently
large to include all points of view nor was it intended to be so because of limitations
of time and resources.
This initiative only sought to test the AHP methodology on a problem that had
previously evaded resolution. The size of the panel was thought to be sufficient to
account for the different populations. However, it was agreed that the work is
exploratory in nature and intended to demonstrate how the method can be used
over a short period of time to arrive at a process that moves the negotiation process
forward.
As mentioned above, at no point in the development and evaluation of the
problem was the process easy. In fact, even the ‘‘purpose’’ was not easily agreed
upon and at several points in the 3 days over which the meetings took place, the
panel readdressed what the undertaking was intended to accomplish. It looked at
the purpose of the project from various perspectives in the hope of finding one that
appeared more promising than others that have been tried. The panel brainstormed
all the issues they could think of that had to be considered in the framework. They
are listed in Table 15.1 below as they were identified by the participants and later
organized into categories with no attempt to eliminate possible duplications.
Listing the issues made it easier to identify the concessions, and to structure the
problem. Taking time to structure the problem in as comprehensive a fashion as
may be feasible, is a crucial first step before attempting to prioritize the relative
importance of its constituent parts that have causal influence on the concessions
and actions to be taken. Needless to say, the structure that emerged in the early
discussion depended on the parties, their knowledge, experience and conditioning.
In a strict sense it was a political rather than a scientific structure. In such a
situation, it was not possible to provide a cultural analysis of the parties’ narrative
and framing of the issues.
The exercise in discussing specific issues sometimes seemed to generate
incompatible perceptions of what can and would be achievable in peace negotiations.
15.4 Implementing the process 209
For example, all the Israelis present were adamant that a one state solution is
impossible to contemplate, while Palestinians all agreed that a solution that does not
grant refugees their internationally recognized rights to return is also impossible to
contemplate. But we do know that historically adamant positions have changed when
circumstances change. For proper application of the AHP methodology, it is
important to include in the structure all factors, including those that some participants
feel are so crucial to their preconceived and predetermined positions, that any con-
cession on those issues seems inconceivable.
In order to develop the necessary measurements for prioritization, we need to
calculate the gains and losses for each concession from each of the parties. The
panel developed a total of eight hierarchies involving benefits and costs and
perceived benefits and costs: four hierarchies for the Israeli group and four hier-
archies for the Palestinian group. The exercise in which the 106 issues were
identified through the process of brainstorming served as a stimulus to the thinking
of the participants to deal with the structuring process. Each of the eight hierar-
chies involves a goal, for example, Israel’s Benefits from Palestinian Concessions,
and a set of criteria that are a subset of the issues relevant to that goal. They are
called criteria in terms of which all the possible concessions that were identified
were evaluated by scoring them one at a time. The criteria that were developed for
these eight models were chosen by each of the Israeli and Palestinian participants
respectively. Because of the volume of issues, we found it necessary in developing
15.4 Implementing the process 211
the hierarchies to select as criteria a subset of the most crucial issues. The overall
goal of each of the corresponding criteria in the four hierarchies involved the
apparent equalization of the ratio of the gains to the losses by each side. Con-
cessions by each party are listed in Table 15.2. We list the concessions which the
212 15 A New Approach to the Middle East Conflict
Priorities
Israeli criteria Israelis benefits from Israelis perception Israelis costs from Israelis perception
Palestinian concessions of Palestinian costs their own concessions of Palestinian gains
from Israelis concessions
Benefits
Control Jerusalem and holy places 0.245
Refugee 0.251
compensation and settlement
The Retributive Function
Priorities
Israeli criteria Israelis benefits from Israelis perception Israelis costs from Israelis perception
Palestinian concessions of Palestinian costs their own concessions of Palestinian gains
from Israelis concessions
Demographic 0.191
Administration 0.021
Civil disorder 0.185
Social 0.095
Unification of Jewish people 0.003
Perceived Palestinian benefits
International recognition 0.002
Member nations 0.017
Recognized borders 0.131
15
Peace 0.014
Independence 0.119
Economic trade 0.013
Law and order 0.085
Pride 0.619
Priorities
Palestinian Criteria Palestinian benefits from Palestinian perception Palestinian costs Palestinian perception
Israelis concessions of Israelis costs from their own of Israel gains from
concessions Palestinian concessions
Benefits
Human rights 0.091
Permanent borders 0.029
Sovereign Palestinian state 0.13
(continued)
A New Approach to the Middle East Conflict
Table 15.3 (continued)
15.5
Priorities
Palestinian Criteria Palestinian benefits from Palestinian perception Palestinian costs Palestinian perception
Israelis concessions of Israelis costs from their own of Israel gains from
concessions Palestinian concessions
Vacating of Israelis from settlement 0.167
Freedom of movement 0.000
Shared water and other 0.239
resources
The Retributive Function
Priorities
Palestinian Criteria Palestinian benefits from Palestinian perception Palestinian costs Palestinian perception
Israelis concessions of Israelis costs from their own of Israel gains from
concessions Palestinian concessions
Social 0.032
Quality of life 0.084
International image 0.044
Social harmony 0.019
Not challenging Israel 0.017
Property rights 0.318
Perceived Israeli Benefits
Peace of mind 0.167
Reduce fear of living 0.041
15
Hence A’s gain from a given concession from B may be described as the
product of A’s benefits and B’s costs (as perceived by A). We have the following
ratios for the two parties A and B: (according to A’s perceptions) A’s ratio:
Gain to A from B’s Concession
A’s Perception
P of B’s Gain from A’s Concession
A’s benefits B’s costs from B’s Concession
¼P
B’s perceived benefits A’s costs from A’s Concession
P
where is the sum over all the benefits obtained by A in the numerator and by B
in the denominator. Hence, given A’s ratio, A’s gain is a product of both the utility
benefit received and the cost to B in providing that benefit as described in the
numerator of the equation. The total gain to A is diminished by the product of the
cost to A in concessions given to B and the perception of the benefit received by
B for A’s concessions in the denominator. A’s benefits and costs are readily
measured by A; however, the costs and gains to B are not readily available to
A and are therefore estimated as perceived by A. A expects to have a gain ratio
greater than one which suggests that the gains to A are greater than the perceived
benefits to B. Likewise, B expects to have a gain ratio greater than one. For
equality in ‘trade’ to be achieved, the two parties should be nearly equal in value,
which suggests that the two gain as much as the perceived benefits to and costs of
concessions to the other. B’s utility is given by the function: (according to B’s
perceptions) B’s ratio:
Gain to B from A’s Concession
B’s Perception
P of A’s Gain from B’s Concession
B’s benefits A’s costs from A’s Concession
¼P
A’s perceived benefits B’s costs from B’s Concession
The measure of equality between the parties in the trade of concessions may be
calculated as the ratio of the two ratios:
A’s Ratio/B’s Ratio = Retributive Gain(Loss) to A.
Where the retributive gain is the amount that A benefits from making B ‘pay’
while a loss is accounted for by the amount that A ‘lost’ in the negotiation process.
Under no circumstance would we expect A to agree to concessions when there is a
perceived loss when A has dominance over B. In the case where A has dominance
over B, the best that B can do is to minimize the disparity in gains.
Table 15.5 Israeli’s concession ratings relative to criteria.
218
development
Drop opposition to trade and Negligible Very low Very low – Negligible – High 0.402 0.435
normal relations w/Israel
Incitement of anti-Israeli Excellent Excellent Excellent – Negligible – Very high 0.853 0.923
sentiment in school
Lobby arab states to allow – – – – – – – 0.000 0.000
Israelis Right to return
Make compromise on the status Excellent Low Excellent – – – Medium 0.653 0.707
of Jerusalem
Denounce and reign in violence Excellent Very low Excellent – High – Excellent 0.795 0.860
Seek assistance for a legitimate Negligible Excellent Excellent – – – Negligible 0.556 0.602
settlement of refugees
(continued)
A New Approach to the Middle East Conflict
Table 15.5 (continued)
15.5
Work cooperatively w/Israel Negligible Very low Excellent – Very high – Negligible 0.677 0.732
to return
Make compromise on – Negligible High – Negligible – 0.418 0.605
the status of
Jerusalem
Denounce & reign in – Excellent High – Negligible – 0.628 0.911
violence
Seek assistance for a – Low Excellent – Negligible – 0.539 0.782
legitimate
settlement of
refugees
Sharing of natural – Medium High – Medium – 0.645 0.935
resources
(continued)
A New Approach to the Middle East Conflict
Table 15.5 (continued)
15.5
Criteria Economic Political Religious Psychological Security International Demographic Administration Civil Social Unification Total Ideals
Concessions inuge disorder of Jewish
people
Abandon the idea – – – Excellent Excellent – Excellent – Excellent – – 0.769 1.000
of Jewish state
Accept two-state – – – Medium Negligible – Negligible – High – – 0.371 0.482
solution
comply with UN – – – Excellent Excellent – Excellent – Excellent – – 0.769 1.000
resolutions
Human rights – – – Negligible Negligible – Negligible – Negligible – – 0.231 0.300
Implementation of – – – Excellent Excellent – Excellent – Excellent – – 0.769 1.000
refugee rights
Palestinian freedom – – – Excellent Excellent – Excellent – Excellent – – 0.769 1.000
of movement
Removal of wall – – – Excellent Excellent – Negligible – Excellent – – 0.635 0.826
and other
barriers
Right to economic – – – Negligible Negligible – Negligible – Negligible – – 0.231 0.300
opportunity
Right to education – – – Negligible Negligible – Negligible – Negligible – – 0.231 0.300
Shared – – – Medium Low – Negligible – Medium – – 0.407 0.529
administration
of resources
(continued)
221
Table 15.5 (continued)
222
Priorities 0.026 0.070 0.002 0.119 0.274 0.014 0.191 0.021 0.185 0.095 0.003
Criteria Economic Political Religious Psychological Security International Demographic Administration Civil Social Unification Total Ideals
Concessions inuge disorder of Jewish
people
Shared control of – – – Negligible Negligible – Negligible – Excellent – – 0.360 0.468
holy places
Shared Jerusalem – – – Excellent High – Negligible – Excellent – – 0.580 0.755
Turnover – – – Excellent Negligible – Negligible – Excellent – – 0.444 0.577
settlement
w/wo
compensation
(d) Israeli’s perceptions of Palestinians’ gains
Priorities 0.002 0.017 0.131 0.014 0.119 0.013 0.085 0.619
Criteria International Member Recognized Peace Independence Economic trade Law and order Pride Total Ideals
15
w/wo compensation
223
Table 15.6 Palestinian concession ratings relative to criteria.
224
barriers
Right to economic opportunity Very – Excellent High – Very high – Low – – 0.703 0.79
high
Right to education Very – Excellent Medium – Medium – Low – – 0.639 0.72
high
Shared administration of Very – Very high Medium – Excellent – Low – – 0.698 0.79
resources high
Shared control of Holy places Medium – High Very high – Medium – Very low – – 0.600 0.68
Shared Jerusalem Medium – High Very High – Medium – Medium – – 0.686 0.77
Turnover settlement w/wo High – Very high Excellent – High – Medium – – 0.748 0.85
compensation
(b) Palestinians’ perceptions of Israelis’ costs
Priorities 0.207 0.037 0.234 0.351 0.086 0.084
Criteria Ending of Change of Ziorist Property Settlement Social Unity based on Total Ideals
Concessions superiority narrative restination evicaution restructuring shared identity
Abandon the Idea of Excellent – Excellent Medium – – 0.688 0.951
Jewish state
(continued)
A New Approach to the Middle East Conflict
Table 15.6 (continued)
15.5
refugee rights
Palestinian freedom of Very High – Low Medium – – 0.550 0.761
movement
Removal of wall and High – High Medium – – 0.599 0.829
other barriers
Right to economic High – Medium Medium – – 0.576 0.797
opportunity
Right to education Very Low – Low Low – – 0.376 0.520
Shared administration Very High – High High – – 0.655 0.907
of resources
Shared control of Holy Medium – Medium Very High – – 0.625 0.865
places
Shared Jerusalem Medium – Medium Very High – – 0.625 0.865
Turnover settlement w/ Very High – Very High Very High – – 0.714 0.988
wo compensation
225
Table 15.6 (continued)
226
demand of right of no
return
Declare against Iranian Medium – Negligible – – – Very low – – – Negligible
nuclear development
Drop opposition to trade & Very high – Medium – – – Low – – – Medium
normal relations w/Israel
Incitement of anti-Israeli Very low – Negligible – – – Very low – – – Negligible
sentiment in school
Lobby Arab states to Allow Very low – Very low – – – Medium – – – Low
Israelis right to return
Make compromise on the Very high – Medium – – – Very high – – – Excellent
status of Jerusalem
Denounce & reign in Excellent – Very Low – – – Low – – – Negligible
violence
Seek assistance for a Medium – Medium – – – Medium – – – Excellent
legitimate settlement of
refugees
Sharing of natural resources Medium – Very High – – – Very high – – – High
Work cooperatively w/Israel High – Low – – – Medium – – – Medium
A New Approach to the Middle East Conflict
Table 15.6 (continued)
15.5
Declare against Iranian Very High – – – – – – – Negligible Negligible Very Low 0.349 0.550
nuclear development
Drop opposition to trade Medium – – – – – – – Very High Medium Very High 0.597 0.942
and normal relations
w/Israel
Incitement of Anti-Israeli Very Low – – – – – – – High Negligible Negligible 0.303 0.479
sentiment in school
Lobby Arab states to Negligible – – – – – – – Low High Very High 0.495 0.782
allow Israelis right to
return
Make Compromise on the High – – – – – – – Medium Very High High 0.595 0.939
status of Jerusalem
(continued)
A New Approach to the Middle East Conflict
Table 15.6 (continued)
15.5
Priorities 0.167 0.041 0.016 0.098 0.034 0.037 0.015 0.021 0.130 0.172 0.270
CriteriaConcessions Piece Reduced Resolution Lawringring Acceptance of Acceptance Social Sharing of Ending of Enhanced Trade Total Ideals
of mind of living of Israel resources Islamic world of Israeli harmony religious apartheid economic with
imagination festivals development region
Denounce & reign in Very High – – – – – – – Excellent High High 0.634 1.000
violence
Seek assistance for a Very High – – – – – – – Very Low Low Medium 0.477 0.753
legitimate settlement
of refugees
Sharing of natural Very High – – – – – – – High Medium Medium 0.563 0.889
The Retributive Function
resources
Work Cooperatively w/ High – – – – – – – High High High 0.591 0.933
Israel
229
Table 15.7 Israeli’s ratios
230
Israeli’s ratios Accept Acceptance Acknowledge Acknowledge Agree to Declare Drop Incitement Lobby Make Denounce Seek Sharing Work
two- of non- Israel’s Israel’s compromise against opposition of Anti- Arab compromise & reign in assistance of cooperatively
state contiguous existence as a existence as to demand of Iranian to trade & Israeli states on the status violence for a natural w/Israel
solution state Jewish state an right of no nuclear normal sentiment to of Jerusalem legitimate resources
independent return development relations in school allow settlement
state w/Israel Israelis of refugees
right to
return
Abandon the idea of 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Jewish state
Accept two-state 1.588 1.451 1.550 1.113 2.214 0.000 0.000 2.060 0.000 0.000 1.749 1.109 1.038 0.000
solution
Comply with UN 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
resolutions
Human rights 7.906 7.273 7.717 5.542 11.022 0.000 3.975 10.251 0.000 4.934 8.706 5.522 5.167 4.923
Implementation of 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
refugee rights
15
Palestinian freedom of 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
movement
Removal of wall and 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
other barriers
Right to economic 7.906 7.273 7.717 5.542 11.022 0.000 3.975 10.251 0.000 4.934 8.706 5.522 5.167 4.923
opportunity
Right to education 7.906 7.273 7.717 5.542 11.022 0.000 3.975 10.251 0.000 4.934 8.706 5.522 5.167 4.923
Shared administration 1.928 1.774 1.882 1.352 2.689 0.000 0.000 2.501 0.000 1.204 2.124 1.347 1.260 1.201
of resources
Shared control of holy 1.753 1.613 1.711 1.229 2.444 0.000 0.000 2.273 0.000 1.094 1.931 1.225 1.146 1.092
places
Shared Jerusalem 0.000 0.000 0.000 0.000 1.314 0.000 0.000 1.222 0.000 0.000 1.038 0.000 1.038 0.000
Turnover settlement w/ 1.234 1.135 1.204 0.000 1.720 0.000 0.000 1.599 0.000 0.000 1.358 0.000 0.000 0.000
wo compensation
A New Approach to the Middle East Conflict
Table 15.8 Palestinian’s ratios
15.5
Palestinian’s Ratios Accept Acceptance Acknowledge Acknowledge Agree to Declare Drop Incitement Lobby Make Denounce Seek Sharing Work
two- of non- Israel’s Israel’s compromise against opposition of Anti- Arab compromise and reign assistance of cooperatively
state contiguous existence as a existence as to demand of Iranian to trade Israeli states on the status in for a natural w/Israel
solution state Jewish state an right of No nuclear and normal sentiment to of Jerusalem violence legitimate resources
independent return development relations in school allow settlement
state w/Israel Israelis of refugees
right to
return
Abandon the idea of 1.060 1.300 1.420 1.057 1.235 3.691 1.175 4.866 1.864 0.000 1.588 1.217 1.054 1.196
Jewish state
Accept two-state 0.000 1.072 1.171 0.000 1.019 3.045 0.000 4.014 1.538 0.000 1.310 1.004 0.000 0.000
The Retributive Function
solution
Comply with UN 1.345 1.350 1.802 1.342 1.567 4.686 1.491 6.177 2.366 1.182 2.016 1.545 1.338 1.518
resolutions
Human rights 1.156 1.418 1.548 1.153 1.347 4.026 1.281 5.306 2.033 1.015 1.732 1.327 1.149 1.304
Implementation of 1.231 1.510 1.649 1.228 1.434 4.287 1.364 5.661 2.166 1.081 1.844 1.413 1.224 1.389
refugee rights
Palestinian freedom of 0.000 1.067 1.166 0.000 1.013 3.029 0.000 3.993 1.530 0.000 1.303 0.000 0.000 0.000
movement
Removal of wall and 0.000 1.197 1.307 0.000 1.137 3.398 1.081 4.479 1.716 0.000 1.462 1.120 0.000 1.101
other barriers
Right to economic 0.000 1.093 1.194 0.000 1.038 3.105 0.000 4.092 1.568 0.000 1.336 1.023 0.000 1.006
opportunity
Right to education 0.000 0.000 0.000 0.000 0.000 2.136 0.000 2.815 1.078 0.000 0.000 0.000 0.000 0.000
Shared administration 1.006 1.234 1.348 1.004 1.172 3.504 1.115 4.619 1.769 0.000 1.508 1.155 1.000 1.135
of resources
Shared control of holy 0.000 0.000 1.022 0.000 0.000 2.667 0.000 3.502 1.342 0.000 1.143 0.000 0.000 0.000
places
Shared Jerusalem 0.000 1.070 1.169 0.000 1.017 3.040 0.000 4.006 1.535 0.000 1.308 1.002 0.000 0.000
Turnover settlement w/ 1.129 1.385 1.513 1.126 1315 3.933 1.251 5.184 1.986 0.000 1.692 1.296 1.123 1.274
wo compensation
231
232 15 A New Approach to the Middle East Conflict
The ratios for the Israelis and the Palestinians based on the benefits and costs,
and perceived benefits and costs are given in Tables 15.7 and 15.8. When a ratio
was less than one the ratio was made equal to zero to signify that a tradeoff was not
acceptable. These ratios are used to make tradeoffs among concessions according
to two criteria: (1) both parties should get as much as possible from the conces-
sion(s), and (2) the gains from the concession(s) for both parties should be as close
as possible. The first criterion is the objective function of a MaxMin problem,
while the second criterion is imposed as a constraint (see Appendix 2).
One of the key takeaways that all participants in the exercise appreciated was that
they learned more about the other party. One of the primary challenges to the
approach turned out to be the same item that created a greater depth in under-
standing: a lack of common definitions. A lack of common definitions challenged
the participants to actively engage in deeper understanding of each other. For
future exercises of this sort though, we suggest that one of the first steps to pursue
15.6 Discussion: The Gain to Loss Ratios of Concessions Made by Both Sides 233
is to define terms and language. For instance, one of the concessions that is offered
is to direct more effort to ‘‘Human Rights’’. How the parties define ‘human rights’,
however, differs greatly. Even though philosophically there should be an easily
determined common definition for ‘‘human rights’’, the reality is that the parties
took different positions on this issue.
The judgment and prioritization process for the concessions was implemented
for each party without knowledge of the other party.
The object is to make the ratios of the two parties close. Each party can by itself
estimate the gain to loss ratio of its opponent and determine if his gain to loss ratio
is much greater than the other party’s gain to loss ratio. That also makes the
negotiations more difficult. The original model sought a solution that matched the
best one-to-one concession. However, given that the best solution was a standoff,
we found that one had to consider trading off bundles of concessions. The role of
the mediator is extremely important in this setting. There are two ways that the
mediator can help to alter the outcome of ratios. Recall that both the numerator and
the denominator of the ratio includes perceptions of the other; in the numerator is
what one party perceives the cost of concessions are to the adversary; whereas the
denominator includes what one party perceives the other party’s benefits. It is
interesting to note that in a retributive conflict one party perceives the costs to the
adversary as a benefit to itself and conversely the gain to the adversary as a loss to
itself. The mediator has a real opportunity to bridge gaps, given the measured
difference between the two parties and their varying perceptions, interpretations
and respect for ‘‘international’’ law.
Our results underline the differences between the Israelis and the Palestinians.
In particular, the findings highlight the value of the Israelis’ concessions as
measured by the Palestinians when compared with the Palestinians’ concessions as
measured by the Israelis through the large differences in ratios. Given this dis-
parity, there is great opportunity by one party to take a leadership role in the
resolution process. Moreover, there is an even greater opportunity for a mediator to
help bridge the gap in the gain-to-loss ratios. By educating both parties on the true
costs and benefits to the adversary, the perceptions are brought more in line with
reality and the score differences minimized. It is possible that external influences
or pressures might be necessary to rationalize the difference in the gain-to-loss
ratios in order to recognize the discrepancies.
Mistrust and the inclination to act retributively prevent people from making all
their concessions at once. To determine the fairest and maximum gain to both
parties from concessions being traded off, we computed gain–loss ratios for each
pair of concessions, one for each party. These gain–loss ratios represent the gain to
one party from the concession made by the other party divided by that party’s loss
from the concession it made. The gain to one party’s concession is obtained as the
234 15 A New Approach to the Middle East Conflict
benefits accrued from the other party’s concession multiplied by the perceived
costs to the party making the concession. The loss to one party’s concession is
obtained as the costs of the concession it made multiplied by the perceived benefits
to the other party. To make the tradeoffs we considered only pairs of concessions
with gain–loss ratios for both parties greater than one. This means that either side
would be reluctant to trade-off a concession in return for another from which its
gain is less than its loss.
The tradeoff process started by attempting to trade-off single concessions with
two objectives in mind: closeness of the gain-to-loss ratios and maximization of
the ratio. If there were no single concessions that could be traded that satisfied both
criteria to ensure fairness, then groups of concessions were considered for tradeoff
to satisfy the same requirements. As a result of this process we obtained
Table 15.9.
The concessions in Table 15.9, numbered as in Table 15.2, are traded off while
maximizing the gain to loss ratio for each side and minimizing the difference
between these ratios to within 1% from each other except for the very last con-
cessions which can only be traded if the 1% constraint is relaxed. All the con-
cessions have been traded off except for the ones related to the resettlement of
refugees (concession 1 for the Israelis from Table 15.2) and the acceptance of
Israel by the Palestinians as a secure, independent and democratic Jewish state
(concessions 3, 6 and 8 for the Palestinians from Table 15.2). These concessions
will need to be addressed as the process is continued.
The outcome shows that with the exception of the two concessions mentioned
above, all the other concessions can be traded off either singly or in groups against
other concessions without violating the constrains previously established, namely
that the gain-loss ratios be not too large and as close to one another as possible.
One might question the advantage of trading off all of the concession identified to
date without addressing the two major issues mentioned above. Prior negotiations
have been hampered by the chaos or confusion caused by trying to address all of
the issues at the same time. What this process permits is eliminating either tem-
porarily or permanently any discussion that would impede attention to the two
major matters that seem the most acrimonious and potentially irreconcilable.
Another reason to take care of the tradable issues first is to give the parties
experience in interacting successfully with one another, thus engendering a spirit
of trust which could be very helpful when the final few but crucially important
issues are considered.
Note, for example that the Israeli concession, ‘‘Shared Administration of
Resources’’ trades off against the Palestinian concession ‘‘Drop Opposition to
Trade and Normal Relations with Israel’’ with a resulting gain–loss ratio for both
sides of about 1.03. Similarly, the remaining concessions are traded off against
those of the other side in groups of two and three. The final two Israeli concessions
in Table 15.3 do not tradeoff against the last Palestinian concession.
It should be noted that the solution to the refugee problem includes removing
the three concessions 3, 6 and 8, all of which relate to this issue. However, only the
single Israeli concession 1 needs to be removed from the present deliberations.
15.7 Equalizing Concession Trade-Offs 235
Tradeoffs of these concessions would require focusing the AHP process only on
these matters with the possible participation of influential outside parties
‘‘The problems we have today cannot be solved by thinking the way we thought
when we created them,’’ said Albert Einstein.
It is possible that either side would be reluctant to accept the outcome of the
tradeoff of bundles of concessions. Optimally these bundles of concessions when
compared both maximize the gain loss ratio for each side and also seek equality of
this ratio for both sides. However, even though the solution may be optimal, there
may still be retributive concern that the other side may be getting more or that they
may have forgotten some important concessions they want or they do not trust
their judgments or cannot completely divest the emotions generated over a long
period of time.
It may be that the Israelis and Palestinians have lived with this conflict for so
long that they do not think an acceptable resolution is possible. People who have
suffered pain for a long time have expectations about what would make them feel
better and a rational solution would not necessarily satisfy those feelings. We all
have a tendency to believe in the mystical, that the hand of God must be allowed to
do its work and miracles do happen. But does that hand act without action by the
parties?
Given the length of time that this controversy has endured, is there a possibility
that the parties at risk would reject an outcome that has embedded long standing
biases and emotions into a solution that considers all conceivable factors and
which produces a recommended outcome which gives both parties an approxi-
mately equivalent set of benefits and costs. The answer may be yes. So of what
value has this process been? Should such reluctance be observed, this will in no
way invalidate the efficacy of this approach.
The approach opens up an avenue of new thinking. Even if people reject the
first effort, it may only need to be modified a bit to make it acceptable or in any
case people may reject the unknown but they live with it long enough for it to
become an accepted and natural way of thinking. People may be less likely to
reject a second or a third revised effort.
While it would be appealing for the parties to implement the recommended
solution, it may not be possible do so without the influence or even coercion of
outside parties such as the United Nations, the United States, the European Union
and Russia. Some of the concessions cannot be implemented without supervision.
236 15 A New Approach to the Middle East Conflict
15.9 Conclusions
Our objective has been to see if the Analytic Hierarchy Process (AHP), a new
approach to group decision-making, could be used productively to move forward the
60 year old debate about solving the perplexing Middle East problem. It was not our
intention to use the process to discover a specific solution to the Israeli–Palestinian
conflict. The AHP provided us with a new way to pursue the dialogue in a context
which uses a quantitatively-oriented approach to attach numerical priorities to the
issues in what has been an emotionally charged conflict. Our purpose has been to
introduce a process which, were it to be used by the actual negotiators, might offer
some new ways of moving forward on the heretofore intractable positions adopted
by the parties.
One may ask how can a process like this add meaning to the plethora of
proposed solutions that have either fallen on deaf ears over 60 years or been
destroyed because of the impossibility of implementation? It is important to state
the idea that the AHP is a supplement and not a replacement for face to face
negotiations. Whenever the process has broken down in the past, there has been no
next step to take. A number of entities have stepped forward to try to jump start the
stalled negotiation. The United States has been the foremost player in this reme-
diation effort. They have tried to determine what would be a fair outcome but to
date there has been no real way to measure which initiative would constitute a fair
and equitable package, because the issues are so varied, complex, interrelated and
affected by extreme emotions. AHP provides an alternative approach by helping
the parties to either think outside the box by themselves or engage in exercises
which force this creative behavior. In its simplest terms AHP would require the
actual negotiator to make judgments in a novel way. The outcome of their judg-
ments could provide an outsider like the United States with some confidence that
an AHP type solution would yield an outcome that is as fair to each side as is
possible with current technology. A third party could then encourage the parties to
consider such a solution with increased confidence that the approach allows, for
example, the United States to act in a neutral position with some confidence that as
fair a solution as possible is being promoted.
It is important to note that the participants in our meetings were knowledgeable,
informed, thoughtful Israelis and Palestinians, who might be able to recommend to
actual negotiators ideas and ways to solve this long-simmering problem. They
were engaged in a simulation—a process to find out if these participants had been,
indeed, the actual negotiators, would they have been able to stay with the nego-
tiations and to reach some productive outcome. Whether these participants were
representative of their respective constituencies is irrelevant, since the results of
their deliberations were never expected to produce a solution to the problem, but
merely to test an approach. The outcomes of the deliberations suggested a number
of important benefits if the approach were actually used by real-life negotiators.
We have attempted to find out how the process would work: would it provide a
modicum of an objective basis to trade off the concessions and help to drain the
15.9 Conclusions 237
have overlooked one or more of these issues and concessions and thus our
structure may be incomplete. That in no way diminishes the effectiveness of our
simulation. However, before beginning to work with combinations of conces-
sions to trade them off, the structure can be trimmed down to include only what
are now known to be the major elements.
2. The outcome in many cases reinforced the conventional wisdom of the par-
ticipants as to what the concrete objectives of each side are and what positions
either side is willing to modify with concessions from the other side or is not
willing to modify regardless of the other side’s concessions. But it was now
possible to measure the gains and losses related to various concessions as
identified by the party that would be providing the concessions as well as their
judgment as to what the benefits and costs were to the other party in providing
their concessions. Obviously the opinion of one side about the costs and ben-
efits to the other side of specific concessions sometimes varied widely from the
other side’s opinion as to the costs and benefits to them of the concessions they
might possibly make. These differences in perception are revealing and they
often led to differences in gain-to-loss ratios as perceived by each party.
3. The AHP process made it possible for the participants to consider a wide
variety of potential tradeoffs, either individually or in bundles. By attaching
quantitative values to the comparisons, a great deal of the emotionality of the
discussion was defused. It became clear that, at least in the case of these
specific participants, certain issues appeared not to be tradable or that the
participants did not know how to trade them. For example, the Israeli need to
have a Jewish state and the Palestinian need to have a satisfactory solution to
the resettlement of refugees appeared to be issues which could not be easily
compromised, if at all. One might surmise that solutions to such issues might
require the involvement of outside parties and that solutions not totally
acceptable to either side might have to be imposed. By comparing costs and
benefits of concessions as viewed by either party and establishing hierarchies
which permitted comparisons of the issues and concessions, some equivalence
of pain or cost and of benefit, either by individual comparisons or in bundles,
might suggest a reasonably objective statement of what a ‘‘fair’’ or equally
painful or equally beneficial outcome might be. Neither side might feel such a
solution would be a ‘‘win’’ for them, an objective each side would prefer to
achieve. They might be convinced, perhaps only by outside parties, that a
solution that would bring peace could only be achieved if each party recognized
through a process such as the AHP or otherwise, that trading off a similarly
beneficial and painful (as objectively measured) set of solutions is the only way
to achieve peace. A further advantage would be that outside parties, such as the
United States or the European Union or the United Nations, etc. could pressure
the parties to settle, using solutions carefully balanced to favor neither side and
exacting compromises from both sides that these outside parties could feel
reasonably comfortable that their impact was objectively measured.
4. Participants were willing to talk about sensitive issues as part of the concession
discussion without feeling threatened by the other party. Though the
15.9 Conclusions 239
possible through the use of the Analytic Hierarchy Process, a chance of resolution
is enhanced. What have the parties got to lose?
Appendix 1
Palestinian Concessions
Lose Revolutionary Cause as Unifying Factor Accountability & Responsibility (Internal & External)
Palestinian Concessions
Accept a Two-State Solution Lobby Arab states to allow both Israelis and Denounce Iranian pursuit of
which includes a Palestinians to have the right to return to their land nuclear arms and support
noncontiguous area - Gaza of origin Israel's efforts to remove the
Acknowledge Israel's Seek assistance for a legitimate settlement of threat
existence as a Jewish State refugees
Israeli Concessions
Encourage equal opportunity for
Abandon the idea of a Share Jerusalem as both a religious
Palestinians to achieve equal
Jewish State & political center w/all parties
economic prosperity
Accept a two-state
solution
Comply w/all applicable Turnover settlements of Jewish Allow the right to have an
United Nations settlers on land claimed by the education that is non-biased and
resolutions Palestinians with or without equally shares historic
compensation backgrounds
Allow the sharing of all
natural resources between Permit Palestinian freedom of
Palestinians and Israelis Comply w/human rights
movement
Israeli Concessions
Abandon the idea of a Share Jerusalem as both a religious & political Encourage equal opportunity for Palestinians to achieve
Jewish State center w/all parties equal economic prosperity
Accept a two-state
solution
Allow the right to have an education that is non-biased
Comply w/all applicable
Turnover settlements of Jewish settlers on land and equally shares historic backgrounds
United Nations
resolutions claimed by the Palestinians with or without
compensation
Allow the sharing of all
natural resources
between Palestinians & Permit Palestinian freedom of movement
Israelis
Israeli Concessions
Abandon the idea of a Jewish Share Jerusalem as both a
State Encourage equal opportunity for Palestinians to achieve equal
religious & political center
economic prosperity
Accept a two-state solution w/all parties
Turnover settlements of
Comply w/all applicable Jewish settlers on land Allow the right to have an education that is non-biased &
United Nations resolutions claimed by the Palestinians equally shares historic backgrounds
with or without compensation
Israeli Concessions
Abandon the idea of a Jewish State Share Jerusalem
as both a Encourage equal opportunity
religious & for Palestinians to achieve
Accept a two-state solution political center equal economic prosperity
w/all parties
Turnover
Comply w/all applicable United Nations resolutions
settlements of
Allow the right to have an
Jewish settlers
education that is non-biased
on land claimed
& equally shares historic
by Palestinians
backgrounds
w/or without
Allow the sharing of all natural resources between Palestinians & Israelis compensation
Economic Religious International Image Property Rights Access to Family Ties Moral
Divisio
Land/Sea Control Social Social Harmony Public Standing International Standing n
Palestinian Concessions
Accept Two-State Solution Agree to compromise on the demand of the right Work cooperatively & in active
of return engagement w/Israel
Accept a Two-State Solution which
includes a noncontiguous area - Gaza
Lobby Arab states to allow both Israelis & Denounce Iranian pursuit of nuclear
Palestinians to have the right to return to their arms & support Israel's efforts to remove
Acknowledge Israel's existence as a land of origin the threat
Jewish State
Reduce Fear
Leveraging Resources Acceptance of Israelis
of Living Sharing of Religious Festivals
Ending of
Apartheid of Sharing of Religious Festivals
Israel Trade within a Region
Palestinian Concessions
Accept Two-State Solution Agree to compromise on the
Work cooperatively & in active engagement w/Israel
demand of the right of return
Accept a Two-State
Solution which includes a
Lobby Arab states to allow both
noncontiguous area - Gaza
Israelis & Palestinians to have Denounce Iranian pursuit of nuclear arms & support
the right to return to their land of Israel's efforts to remove the threat
Acknowledge Israel's origin
existence as a Jewish State
Appendix 2:
How to Select Concessions from One Party to Match Concessions
from Another Party
To decide how to match the concessions of one party with the concessions from
another party we need to first create all possible concession bundles for both parties.
A concession bundle is a set of individual concessions. The parties can then trade
concession bundles. The problem is that there are many possible concession bun-
dles even when the parties in conflict have a moderately small number of possible
concessions. For example if one party had 13 concessions and another had 14
concessions, there are 8,191 and 16,383 possible concession bundles, respectively.
Since we need to match a bundle of one party with all other possible bundles of the
other party, to determine which concession bundle is more advantageous, we need
to solve 7,563 matching problems for one party and 14,787 problems for the other
party. Were we to do it all at once, then the problem would be even more difficult to
solve, because the problem would involve 8,191 9 16,383 = 134,217,728 vari-
ables. A possible solution is to divide the concessions into groups such as short,
medium and long term sets and then form the bundles.
Let CA and CB be the set of concession bundles of two parties A and B in a
conflict. Let ci ðkÞ be the ith concession bundle of party k. Let pði; Ajj; BÞ be the ratio
gain from the ith concession bundle of party A when party B offers the jth con-
cession. Let qðj; Bji; AÞ be the ratio gain from the jth concession bundle of party
B when party A offers the ith concession. Let xij be a binary variable where xij ¼ 1 if
the ith concession bundle of A is matched with the jth concession bundle of B.
Concession bundles from one party can be paired withP concessions
P bundles of
the other party. Thus, the total gain of party A is given by pði; Ajj; BÞxij and
i2CA j2CB
P P
the total gain of party B is given by qðj; Bji; AÞxij : To balance both gains and
i2CA j2CB
provide both parties with the maximum gain, we solve a MaxMin problem, i.e., a
maximization
P P P P function is an arbitrary variable x0 such that
model whose objective
pði; Ajj; BÞxij x0 and qðj; Bji; AÞxij x0 : If all the concessions are
i2CA j2CB i2CA j2CB
P P
matched then xij ¼ 1 and xij ¼ 1: If only a subset of SA CA is matched
i2CA j2CB
P P
with CB then xij ¼ 1; for i 2 SA and xij 1: If only a subset SA CA is
j2CB i2CA
P P
matched with a subset SB CB ; then xij 1 for i 2 SA and xij 1 for
j2CB i2CA
j 2 SB : Thus the more general problem is given by:
Appendix 2: How to Select Concessions from One Party 247
Max x0
s:t:; X X
pði; Ajj; BÞxij x0
XA j2C
i2C XB
qðj; Bji; AÞxij x0
i2CA j2CB
X X XX
pði; Ajj; BÞxij qðj; Bji; AÞxij e
i2C j2C
XA B i2CA j2CB
xij 1; i 2 SA
XB
j2C
xij ¼ 0; i 62 SA
XB
j2C
xij 1; j 2 SB
X
i2C A
xij ¼ 0; j 62 SB
i2CA
xij ¼ 0; 1; i 2 CA and j 2 CB
Bibliography
1. Saaty TL (1989) International center for conflict resolution, defense analysis, vol 5(1).
Brassey’s Defense Publishers Ltd., Great Britain, pp 80–82
2. Saaty TL (2010) Principia mathematica decernendi. RWS Publications, PA
3. Saaty TL, Vargas LG (2006) Decision making with the analytic network process: economic,
political, social and technological applications with benefits, opportunities, costs and risks.
Springer’s International Series
4. Zoffer et al (2008) Synthesis of complex criteria decision making: a case towards a consensus
agreement for a middle eastmiddle east conflict resolution. Group Decis Negotiation 17:363–385
Chapter 16
Legalization of Euthanasia
16.1 Introduction
The Oxford English Dictionary (2nd edition, 1989; online version June 2011)
provides the following definitions of euthanasia:
A gentle and easy death… The means of bringing about a gentle and easy death…
In recent use: The action of inducing a gentle and easy death. Used esp. with reference to a
proposal that the law should sanction the putting painlessly to death of those suffering
from incurable and extremely painful diseases.
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 249
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_16, Springer Science+Business Media New York 2012
250 16 Legalization of Euthanasia
The legalization of euthanasia is a highly charged political issue. Like many other
political issues, its fate is determined by key players—politicians, the general
public, and strong lobbying groups—and their view of the major issues within the
dilemma. The model used to answer the question of whether or not euthanasia
should be legalized is the benefit—cost analytic hierarchy. The risks associated
with the legalization of euthanasia are incorporated in the cost hierarchy because,
in a life-critical situation, all risks inevitably become costs. Because we are not at
liberty to gamble with human life, we must assume the worst case scenario and
take all risks to be costs. The key players surrounding this issue are the criteria and
the key issues they face are the subcriteria, followed by the alternatives.
support the measures. The AMA is a powerful political lobbyist, and therefore
placed third in overall ranking of key players.
Hemlock Society—In our model, the Hemlock Society represents all right-to-die
groups. The Hemlock Society has been in existence for over twenty years and its
members are very vocal in support of all forms of euthanasia. Although the group
placed last in the rankings, its influence should not be ignored. An organization
like the Hemlock Society placed the assisted suicide bill on the ballot in Oregon in
1996.
Economic—As with any medical procedure, assisted suicide and active euthanasia
will involve some financial costs. These may include the cost of a prescribed drug
overdose for assisted suicide or fees paid to the attending physician for active
euthanasia. The increase in medical malpractice liability should also be consid-
ered, and although current malpractice insurance is likely to cover euthanasia suits
[10], it is also likely that the legalization of euthanasia will drive up the cost of
malpractice insurance to doctors and hospitals.
Moral—Refers to the notion of whether it is morally right to take one’s own life
and for someone else to assist this endeavor. This criterion weighted heavily with
the general population and with religious groups.
Legal—This subcriterion looks at the possible lawsuits that medical profession
could face for potential abuses of the system, and for an unwillingness to partic-
ipate in assisted suicide. In addition, physicians may face liability in cases of
assisted suicide where the drug overdose did not cause the patient to die. As many
as a quarter of those who try to kill themselves with a prescribed overdose could
linger for hours or days before they died [11]. The legal costs also refer to an
increase in legal activity in the government due to necessary legislation to pass this
bill and regulate its use.
Patient Concerns—The major cost in patient concerns in assisted suicide is the
risk of failed procedure. Because doctors are not trained in terminating life, their
‘‘fatal’’ prescriptions come with no guarantees. Patients who take an insufficient
dose could suffer increased pain and incapacitation. In addition, there is the risk
that the patient is not mentally fit to make the decision to seek assisted suicide or
active euthanasia. Many people who seek to commit suicide are motivated by
depression, mental illness, or emotional distress, rather than by a rational evalu-
ation of the situation and subsequent logical decision. A final cost is the families’
potential difficulties in accepting this decision.
16.2.2.3 Alternatives
16.3 Results
Table 16.1 summarizes the points of view of the different constituencies. Note that
Politicians, the general population and the AMA think that Assisted Suicide has the
most benefits while religious groups and the Hemlock society find themselves at
opposite extremes. In terms of costs, everybody but the Hemlock society think that
active euthanasia is most costly
256 16 Legalization of Euthanasia
The final result is to keep the status quo, although assisted suicide was a close
second and active euthanasia finished a distant third. The final decision is not
surprising and reflects people’s unwillingness ‘‘to rock the boat’’. Most people
would rather avoid making such a choice because of the delicacy of the issue. No
one really wants to face death and no one wants to play God. We believe the
moral/ethical issue played the biggest role of all the subcriteria in the final decision
because it is the most ambiguous. Costs and benefits values can easily be measured
16.3 Results 257
for economic and legal issues, while in the cases of moral/ethical and patient
concerns, value is very much subjective. Finally, the country seems to be heading
in conservative direction, and maintaining the status quo confirms that trend.
To determine the strength and resilience of our model, we must conduct sensitivity
tests. Although this is an issue of national importance, we expect that different
regions within the U.S. would impose varying degrees of latitude on euthanasia. In
our sensitivity analysis, we looked at Western states, Southern states, and the
Pittsburgh area. Our results are given in Table 16.2, 16.3 and 16.4.
16.4 Sensitivity Analysis 259
In our analysis for the West, we incorporated the growing sentiments against the
government and in favor of individual rights among the citizens by decreasing the
influence of politicians from 28.50 to 18.9%, and increasing the influence for
the Hemlock Society by nine percentage points. Consequently, assisted suicide won
by a very slight margin. We found this to be strongly supported by the situation in
Oregon, where the bill legalizing assisted suicide was passed by a 51–49% vote.
Citizens in the South are very religious and are more conservative than the rest
of the population. Therefore, we increased the percentage for the religious groups
from 6.50 to 23.9%, This confirmed our initial results in that the status quo
received a higher ratio then assisted suicide and active euthanasia. The major
difference is that people in the south consider assisted suicide to be worse than
active euthanasia.
260 16 Legalization of Euthanasia
Pittsburgh has a large catholic population and is renowned for its medical
facilities. As a result, we increased the religious factor to 17% and increased the
AMA’s influence to 17% as well. This led to a final decision to keep the status quo
by a large margin. This is not too surprising seeing that the religious groups are
categorically against assisted suicide and active euthanasia.
In all three cases, active euthanasia was never a consideration because most
citizens, regardless of geographic location, do not look favorably on a physician
taking life instead of sustaining it.
262 16 Legalization of Euthanasia
16.5 Conclusions
The analysis done using the AHP, suggests that the Status Quo should be main-
tained. The Status Quo states that if a patient requests to discontinue medical
services, then a doctor must abide by his or her wishes. The sensitivity analysis
confirmed this decision except in the West, where legalization of euthanasia has
advanced much further than any other region. The end results confirm a conser-
vative trend which has been sweeping the country, and also indicates people’s
unwillingness to take life into their own hands. The issue of assisted suicide and
euthanasia will not be resolved overnight, and will continue to divide people along
geographic, racial, and ethical lines.
Bibliography
1. Drinan RF (1994) The law and assisted suicide. America. 4 June 1994, p 7
2. Medicine’s Position is Both Pivotal and Precarious in Assisted-Suicide Debate,’’ JAMA.
February i, 1995, p 363
3. lbid
4. Life (1995) Death, and the Pope, Newsweek. April 10, 1995, p 59
5. States Weigh Assisted Suicide (1995) American Medical News. February 27, 1995, p 24
6. Pallone Nathaniel J (1992) Costs and benefits of medicide. Societv July/August 1992, p 34
7. Medicine’s Position is Both Pivotal and Precarious in Assisted-suicide Debate, JAMA.
February 1, 1995, p 364
8. Drinan, Robert F. ‘‘The Law and Assisted Suicide,’’ America. June 4, 1994, p 6
9. Oregon Doctors Fear Fallout From Assisted Suicide, American Medical News. January 23,
1995, p 21
10. Ibid
11. Ibid, p 26
Chapter 17
How Should Congress Address
the Medicare Crisis?
17.1 Introduction
The Medicare program was initiated in 1965 when the federal Social Security Act
of 1965 was passed. Title 18 of this act established a two section provision for the
Medicare program. Part A, which provides health benefits to its beneficiaries,
protects them against hospital related costs. This provision is financed through a
2.9% Social Security payroll tax. Part B provides supplemental medical insurance
benefits to protect enrollees against the costs of physician services, supplies, tests
and some home health services. This provision is financed through voluntary
premiums and matched by funds from general revenues.
Medicare has been, perhaps, the most successful of America’s social programs.
Almost all Americans sixty-five and older obtain health insurance through the
Medicare program. In order to keep this program successful, it has been modified
nine times through provisions such as the Social Security Act of 1972, which
established Professional Standards Review Organizations to monitor necessity and
quality of services, and the Omnibus Budget Reconciliation Act which removed
co-payments for Part B services and limits on home health care visitations.
While Medicare has been a success to the approximately 47.7 million people to
whom it directly provides service, it also has an impact on health care providers. In
Pennsylvania, Medicare revenues account for about 57% of total hospital days.
Additionally, some rural hospitals in Central Pennsylvania have an even greater
dependence on Medicare funds. Lastly, Medicare provides almost all the revenue
received by home health agencies, hospices and renal dialysis facilities. For these
reasons, Medicare is far more important to the health care industry as a whole than
to the elderly alone. Due to the number of people who are impacted by the
Medicare program, Congress faces a difficult political and public issue. The reason
that Congress must address this issue is because the Medicare trust fund, which
reimburses providers for services delivered to Medicare recipients, is being
depleted on a daily basis.
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 263
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_17, Ó Springer Science+Business Media New York 2012
264 17 How Should Congress Address the Medicare Crisis?
Fig. 17.1 U.S. life expectancy at birth 1940–2007. Source: National Center for Health Statistics
(2010) Deaths: Final Data for 2007. Hyattsville MD. Access at http://www.cdc.gov/NCHS/data/
nvsr/nvsr58/nvsr58_19.pdf
The ability of Medicare to remain successful is currently being tested. The main
reason for this crisis is that Medicare expenses are expected to rise more rapidly
than the revenue generated by payroll taxes. The reason for these increases is the
combination of the high cost of health care and an increasingly aged population. In
addition to the increase in age of our population, the elderly population has an
increasing life expectancy (see Fig. 17.1).
Currently, the eighty five and over age group is the fastest growing aspect of the
population in the United States. To make matters worse for Medicare financing,
this group also consumes the most medical care per capita.
According to the ‘‘2010 Annual Report of the Boards of Trustees of the Federal
Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds’’ the
Medicare Trustees are required to test annually whether general revenues will
finance 45% or more of total Medicare spending in any of the next 7 years. In 2010,
for the fifth year in a row, the Trustees projected that general revenues will exceed
45% of total spending within a 7 year timeframe (in 2010), prompting them to issue a
‘‘Medicare funding warning.’’ However, general revenue is projected to fall below
the 45% level in 2011 and not reach that level again until 2022 (see Fig. 17.2).
The number of Medicare beneficiaries is expected to increase. In addition, the
baby boomers will begin to tremendously increase the number of Medicare
enrollees in the year 2010. The problem of providing coverage for this increase in
enrollees is compounded by legislative initiatives to reduce the federal financing of
Medicare (see Fig. 17.3).
Given this increase in membership and reduction in funding, a crisis is on the
horizon for the Medicare program. According to projections by the Kaiser Family
Foundation based on data from the 2009 and 2010 Annual Report of the Boards of
Trustees of the Federal Hospital Insurance and Federal Supplementary Medical
Insurance Trust Funds, the Part A Trust Fund is projected to be depleted by 2029. This
is largely due to reductions in the growth rate of Medicare spending as a result of
provisions in the 2010 health care reform law, as well as a provision to increase the
payroll tax paid by higher-income people. As a result, the Part A Trust Fund is pro-
jected to have a positive asset balance of $317 billion at the end of 2019 (see Fig. 17.4).
17.1 Introduction 265
Fig. 17.3 Income and expenditures of the medicare part A trust fund
Because this issue continues to cause political uproars from the health care
industry and political action committees such as the American Association of
Retired People, legislative action will be required to resolve this issue.
Before we used the AHP decision making model we conducted research and
utilized our own knowledge of the health care industry to generate potential
alternatives which Congress could utilize in resolving the dilemma facing
Medicare. During this process, we generated eight possible alternatives. They are:
266 17 How Should Congress Address the Medicare Crisis?
Fig. 17.4 2009 and 2010 projections of the medicare part A trust fund balance
Costs (0.160): We derived seven potential costs which affect most alternatives to
solving the Medicare crisis. The seven costs and their weights are:
1. National Debt (0.031): Increases in the national debt, through increases in
Medicare expenditures, are highly undesirable. Since Medicare is a very costly
program, and we want to control expenses, we assigned a very low weight to it.
2. Political (0.045): Legislators have realized the political costs of making a
wrong decision concerning Medicare for many years. Now that Medicare
reform is urgent, legislators will want to minimize their potential costs of any
decision.
3. Inflation (0.068): Congress must attempt to keep the growth of this program
parallel to the inflation rate. Currently, the increase in Medicare expenditures
not only exceeds the rate of inflation, but it is increasing at a rate greater than
that of general inflation. For this reason, cost reduction efforts are desirable.
The final four costs and their weights are:
4. Health status changes in the elderly (0.104).
5. Administrative costs associated with administering the program (0.159)
6. Problems with the reduction in access to medical care (0.240).
7. Economic welfare losses associated with changes in individual and government
expenditures (0.354).
Risks (0.227): We generated five different possible risks for potential alternatives.
Once again, please note that the biggest risks possess the smallest weights. These
risks and their weights are:
1. Bankruptcy (0.062): The largest and most significant risk to the program is
bankruptcy. Due to obvious health and industry issues, no one wants this
program to dissolve because of a lack of available funds. Since the current level
of Medicare expenditures already exceeded income in 2010 (see Fig. 17.3), we
gave this factor significant weight.
17.3 Analysis of Benefits, Risks, Costs and Opportunities Components 269
2. Political (0.097): Medicare has many political risks for legislators. Due to the
immense lobbying power of the American Association of Retired People, the
American Medical Association and other interest groups affected by Medicare,
legislators have taken a ‘‘hands-off’ approach whenever the Medicare debate
develops. The fear of losing votes during their next election is very real, so this
risk carries notable weight.
3. Rationing of Services (0.160): If Medicare reimbursements continue to be
reduced, the services of providers may need to be rationed among enrollees.
The reduction of health care services to the elderly is a significant risk, and is
present in most options. Rationing is generally viewed as unacceptable.
The final two risks and their weights were:
4. Increasing the total number of uninsured (0.263).
5. The achievement of significant dollar savings (0.419).
Although ethical behavior was not one of the four basic categories used in eval-
uation, the topic does deserve mention. Regardless of the results AHP helps to
derive, it would not be right for people who paid Medicare payroll taxes
throughout their working careers to not receive the benefit of insurance coverage.
A potential resolution to this issue would be that the Medicare payroll tax be
structured and viewed as a type of insurance for medical care if people do not have
the means to buy it themselves when they are old.
Based on their decisions, members of Congress face the threat of losing
campaign donations. Legislators must try to be ethical in their decision making and
not let this threat sway their judgment. Additionally, hidden agendas and
bureaucracy must not take precedence over such a major issue. Legislators must
work together for the betterment of our country and not the betterment of their
career and political party.
17.5 Results
To rate the alternatives according to their benefits, opportunities, costs and risks,
we first constructed rating scales given in Table 17.1. The description of each of
the intensities in the scales is given in Table 17.2.
In Table 17.3 each of the alternatives is rated according to their benefit,
opportunity, cost and risk level. Each of the intensities has a numerical value in
Table 17.1. The resulting values are transformed into an ideal scale, i.e., each
270 17 How Should Congress Address the Medicare Crisis?
entry is divided by the largest value of the corresponding scale. The result is given
in Table 17.4.
Next we computed the short term (BO/CR) and the long term (bB ? oO –
cC - rR) value of the alternatives, where
and
where b, o, c and r are the weights of the benefits, opportunities, costs and risks,
respectively.
17.5 Results 271
17.6 Conclusions
We found the results to be very interesting. The AHP helped us to decide that in
the short term Congress should mandate that all Medicare beneficiaries be enrolled
into managed care health insurance plans. This result is favorable to us since the
trend in health care is to change the financing and delivery of services from a
treatment oriented system into a model of care management. This trend is espe-
cially evident in the Medicare market which is rapidly embracing the managed
care methodology. Also, we are aware of the benefits of managed care and believe
that this model can be effective.
Mandated Managed Care enrollment received very favorable weights in the
AHP model due to the fact that this measure can transfer risks to other entities,
achieve expenditure reductions, manage the care of beneficiaries and encounter
only some of the problems with access. All of these components were very
favorable in our analysis.
The surprise in our results was the long term alternative. We never imagined
that the AHP would help us decide that instituting a National Health Insurance
program would be a serious alternative. After reviewing our analysis, we realized
the major reason this alternative came in second was that national health insurance
provides universal health coverage to all individuals. The desire to have as many
people covered as possible was the number one factor (carrying the most weight)
of the number one component in our hierarchy. For this reason, this measure
received a higher rating than we expected, although the negative aspects of
national insurance, such as the high cost, kept it from being the top choice in the
short term.
17.6 Conclusions 273
The costs, risks, benefits and opportunities were all ranked with the highest
value being the most beneficial or detrimental. In analyzing the results on the
attached spreadsheet, it is readily evident that mandating the use of Managed Care
for all Medicare subscribers has the highest benefit to cost ratio (13.58 ratio).
Although other alternatives had higher ratings of benefits and opportunities,
the low cost and risk of managed care was the biggest factor in it being the
recommended choice. Intuitively, this choice seems the most appropriate of the
alternatives offered.
In the short term, there are two results which AHP presents that are surprising.
First, AHP suggests that the second best option is for Congress to Discontinue
Federal Funding of the Medicare program (2.45 ratio). This option is most
undesirable because it would result in the end of the program (which we are trying
to avoid) and millions of individuals having no health insurance coverage. The
consequences of this, both from an economic and health standpoint, would be
significant and disastrous. The second surprise was that an alternative we con-
sidered as potentially successful, Enacting Means Testing, was ranked as one of
the least desirable options (0.33 ratio). This option was attractive to us because it
would reduce the number of individuals on the Medicare program by removing
those who have the means to obtain private insurance and continuing to serve those
in the most need. Although we do not view this as the answer to the whole
Medicare problem, it would be a step in the right direction. However, we saw the
reduction of members and potential growth of private insurance as relatively small
benefits and consequently assigned a low benefit and opportunity rating.
In the long term, Managed Care was second to a National Health Insurance
Program. What make the National Health Insurance Program unattractive in the
short term is the costs associated with it.
Clearly, the issue of the Medicare crisis has many factors which influence the
course of action which Congress will take. Even if an exact answer to the problem
is not obtained, the process of utilizing AHP will help clarify the issues and
priorities pertinent to achieving resolution. With careful preparation, Congress
would be well served through the use of the AHP in identifying and evaluating
favorable options to consider.
Chapter 18
Ethics in International Business
18.1 Introduction
Ethics is the science of judging specifically human ends and the relationship of
means to those ends [1, p. 2]. It is essentially the art of controlling means so that
they will serve specifically human ends. Business ethics is concerned primarily
with the relationship of business goals and techniques to specifically human ends.
It studies the impact of acts on the good of the individual, the firm, the business
community, and society as a whole. While it does not concentrate on the obli-
gations which a person has as a private individual and a citizen, these enter in since
the business person is all three of these people in one. This means that business
ethics studies the special obligations which a person and a citizen accepts when he
or she becomes a part of the business world. Today’s business world is one of ever
increasing competition where there is great pressure on businesses to maintain
increasing growth rates and profit margins to continually drive the company/s
stock price upward. As international competition has dramatically risen in recent
years there has been unprecedented changes that have been forced upon the
workforce and normal business operations.
As the Japanese and European Communities continue to expand their economic
influence, the United States will be increasingly forced to operate internationally to
maintain increasing profits from period to period (quarter to quarter). These
practices tend to focus on short-term returns to limit any possibilities of a
downward trend in the stock price of a manager/s specific company. As the
importance of profitability in the short-run increases, pressure will be focused on
management to behave in an unethical manner in order to compete favorably in the
international arena.
Ethics have always been a key issue for businesses, but as international
expansion continues, temptations will be prevalent to cut corners to achieve
increasing profits. Many foreign countries legally permit unethical practices and
foster a business environment where corruption runs rampant. For example,
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 275
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_18, Ó Springer Science+Business Media New York 2012
276 18 Ethics in International Business
Germany, U.K., and Denmark actually allow incentive payments and term them
tax deductible business expenses [2, p. 80].
This AHP model helps to evaluate the choices the management of each com-
pany will have to make when operating in an international business environment in
regards to ethical or unethical behavior (Fig. 18.1).
Even though every company will publicly admit to only ethical actions within their
business, many companies behave in unethical manners behind closed doors. The
model is based on the viewpoint of a manager within the ranks, and not the
company in general. The first level of the hierarchy breaks down two types of
views on the corporate world—the pragmatic and the moral points of view.
The pragmatic, or practical, viewpoint is the view of the manager who wants to
meet his profit targets regardless of the actions involved or their implications. The
moral, or ethical viewpoint, is the view of the manager who goes about achieving
his or her goals within certain absolute limitations on behavior. The pragmatic
viewpoint is split into four subcriteria; importance of profits, pressure from upper
management to achieve profit goals, drawbacks of acting in an ethical manner, and
society’s norms and regulations. The moral viewpoint is also split into four criteria
that include the ethical teachings of the past, the codes of conduct of the company,
the benefits of acting ethically, and society’s norms and regulations. The synthesis
18.2 The Analytic Hierarchy Model 277
of these criteria with the judgments on priorities of each one will then result in the
final choice of whether to behave ethically or unethically.
Figure 18.2 shows the result of prioritizing the different components of the model
in Fig. 18.1. According to this model 70% of the behavior can be attributed to
Profits (Co. $—0.191; bonuses—0.068), Management Pressure (promotion—
0.095), Societal Moral Pressure (culture—0.129; regulation—0.106) and Teach-
ings (family influence—0.063; work experience—0.063).
In Table 18.1 the ethical/unethical behavior is evaluated in terms of the criteria.
Note that unethical behavior seems to dominate 55/45. Were one to consider just
the main criteria that capture 70% of the priorities we can see (Table 18.2) that
unethical behavior is even more dominant 58/42.
278 18 Ethics in International Business
0.447 0.553
0.419 0.581
18.4 Conclusions 279
18.4 Conclusions
The Analytic Hierarchy Process model suggests that there is a tendency to behave
in an unethical manner when operating in an international business environment.
By doing a sensitivity analysis, we concluded that the less pragmatic a manager
becomes, the greater the chance is acting in an ethical manner. The results of this
study show that, managers have a tendency to act in an unethical manner about
55% of the time, as compared to only operating in an ethical manner about 45% of
the time. This may exemplify the decreasing values of today’s society as short-
term profits become the top priority of the corporate world. Society has never been,
and will never be, completely ethical but it appears that as international compe-
tition increases the ethical standards of businesses are decreasing.
Bibliography
1. Garrett TM, Klosnoski RJ (1986) Business ethics. Prentice-Hall, Inc., Englewood Cliffs
2. Janavaras BJ (1988) International business: introduction and essentials. Business Publications,
Inc., Texas
3. Solomon RC (1983) Above the bottom line: an introduction to business ethics. Harcourt Brace
Jovanovich, Inc., USA
4. Wirtenberger HJ (1962) Morality and Business. Loyola University Press, Chicago
Chapter 19
Abortion and the States: How will
the Supreme Court Rule on the Upcoming
Pennsylvania Abortion Issue?
19.1 Introduction
In the Summer of 1992 the Supreme Court of the United States was supposed to
rule on a controversial Pennsylvania statute [10] restricting the rights of women in
obtaining an abortion. Included in this statute are provisions requiring that doctors
provide women with state-prescribed information about pregnancy and abortion,
that the procedure be delayed 24 h after the recitation and that husbands be
notified prior to the procedure. The lower court upheld the first two provisions but
declared unconstitutional the husband notification requirement.
On January 22, 1973, the United States Supreme Court, in a landmark decision,
ruled on the constitutionality of abortion by handing down judgments in two ‘‘test’’
cases: Roe v. Wade [12] and Doe v. Bolton [8]. The court replaced two states’
statutes by declaring their restrictions on abortion unconstitutional. The court
replaced the invalidated statutes with a uniform system clearly identifying the
stages of pregnancy in trimesters and the ‘‘legal’’ enforcement one can expect
during each trimester.
The Court used the ‘‘strict scrutiny’’ standard which applies to only a handful of
constitutional rights that the Supreme Court has labeled ‘‘fundamental’’. They
determined that the right to privacy, whether in the Fourteenth Amendment’s
definition of personal liberty, restricting government intrusion, or the Ninth
Amendment’s consignment of rights to the people, is broad enough to include the
right to have an abortion. In denying a woman the choice of terminating an
unwanted pregnancy, the state would create a hardship for the pregnant woman
that could result in psychological harm [9, p. 11].
These rulings have come under close scrutiny since conservatives took office in
1980. Presidents Reagan and Bush have collectively appointed seven of the nine
Supreme Court justices. Since the term of the president is four years and the term
of a Supreme Court jurist is for life or until he or she voluntarily resigns, it is
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 281
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_19, Springer Science+Business Media New York 2012
282 19 Abortion and the States
19.2 Subcriteria
Services which gave states more freedom to regulate abortions throughout the
entire pregnancy, and the City of Akron v. Akron Center for Reproductive Health
[5] which required parental consent, informed consent, a 24 h waiting period and
the proper disposal of fetal remains.
Moral issues: We define moral issues as those issues deemed important by the
Pro-Life movement. Included is the constitutional right of the fetus and the belief
that abortion is murder and should be illegal.
Political issues: To make the decision-making process easier, we loosely defined
political issues as conservatism. Although, not always true, we determined that
conservatives are much more pro-life than liberals. Along the same vein we linked
Republicans with conservatism and Democrats with liberalism.
Biological issues: Our broad definition of biological issues is viability. Viability is
a medical concept that specifies a certain time within the gestation cycle when the
fetus is capable of independent survival outside the mother’s womb. This has been
a difficult issue for the court to deal with because it is very closely related to the
issue of ‘‘when life begins.’’ In 1973, the court concluded [9, p. 1]:
we need not resolve the difficult question of when life begins, when those trained in the
respective fields of medicine, philosophy and theology are unable to arrive at any con-
sensus. The judiciary, at this point in the development of man’s knowledge, is not in a
position to speculate as to the answer.
Also, in Akron v. Akron Center for Reproductive Health the Court ruled [5,
p. 491]:
A State may not adopt one theory of when life begins to justify its regulation of abortion.
The Court has thus far ruled that viability should be determined by the attending
physician. The current problem is that modern medicine has allowed for much
earlier viability than existed in 1973. Current medicine has allowed pre-third
trimester fetuses to survive outside the womb. This is an issue which has not yet,
but will likely, be dealt with by the courts.
19.3 Alternatives
We chose three alternatives that we felt were most likely to occur in the June
ruling. In the past, including the very important case of Webster v. Reproductive
Health Services, the Court has been averse to upholding or overruling Roe v.
Wade. Whether or not this will be the case in the Pennsylvania ruling can only be
determined by the Court.
First, we defined OVERTURN as disallowing abortion as a legal right in the United
States. The court would rule that the right to an abortion is no longer a funda-
mental right protected by the Constitution. Basically, abortion would be legal only
when the mother’s life is in danger.
284 19 Abortion and the States
19.4 Criteria
As stated previously, we used as our criteria the nine justices currently sitting on
the Supreme Court. Since each justice is allowed one ‘‘vote,’’ we gave each equal
weight. We used many experts’ opinions from books and law journals to determine
how each justice will weight each criterion and how each subcriterion will affect
the alternative selected. Below we have outlined the major reasons for determining
how each justice will vote:
William Rehnquist: As the most influential of the nine justices, Rehnquist’s
opinion is very important. Prior to his appointment as Chief Justice by President
Reagan, Rehnquist ruled on many abortion cases as a regular jurist. He was one of
two dissenting jurists in the Roe v. Wade decision. For this reason, we decided that
PRECEDENT was not held in high regard by Justice Rehnquist, but that he would give
it some weight (0.134) because of the recent cases limiting rights in which he
concurred. We determined Rehnquist to be a conservative, and, therefore, we
assumed that he would rank POLITICAL ISSUES about the same as precedent (0.118).
In the area of WOMEN’S ISSUES we assumed that he would assign it a very low
ranking (0.055); in fact, we ranked it lower than any of the criteria based on his
ruling in Roe v. Wade. He opined [9, p. 15]:
The fact that a majority of states have had restrictions on abortions for at least a century is
a strong indication that the asserted right to an abortion is not so rooted in traditions and
conscience of our people as to be ranked fundamental. Even today, when society’s views
are changing on abortion, the very existence of debates on the issue is evidence that the
right to an abortion is not so universally accepted as one would have us believe.
In the area of BIOLOGICAL ISSUES, we thought that he would rank it low (0.086)
because of Webster v. Reproductive Health Services in which he ruled that leg-
islative determinations of when life begins are not per se unconstitutional. He
ruled, in essence, that the Court permits a state to determine when life begins, but
maintained that such determinations are not enough to save a statute if it is
unconstitutional for other reasons [4, p. 513]. Finally, we determined that Rehn-
quist would weight the area of MORAL ISSUES very high (0.608) based on his earlier
rulings, his belief that the right to an abortion is not a fundamental individual right,
and his belief that fetuses should have individual rights.
19.4 Criteria 285
Byron White: In most cases which we have researched related to abortion rights,
Justice White has ruled very closely with Justice Rehnquist. For this reason, and
because of his written opinions, we have ranked White very closely with Rehn-
quist. In the area of WOMEN’S ISSUES, we ranked White slightly lower than Rehn-
quist (0.045). And we weighted MORAL ISSUES very high (0.529). These criteria
were ranked based on his opinion in the Doe v. Bolton case [9, p. 28]:
In a sensitive area such as this, involving as it does issues over which reasonable men may
easily and heatedly differ, I cannot accept the Court’s exercise of its clear power of choice
by interposing a constitutional barrier to states efforts to protect human life and by
investing mothers and doctors with the constitutionally protected right to exterminate it.
For the PRECEDENT criterion, we gave it a low ranking (0.097) for the same
reason we did with Rehnquist: because he dissented in Roe v. Wade. For POLITICAL
ISSUES, we rated it low (0.091) because, although he has voted conservative, he was
appointed by a Democratic President. Finally, for BIOLOGICAL ISSUES, we gave it a
fairly high ranking (0.237) because of his belief that life begins at conception and,
therefore, fetuses have rights that should be protected.
Antonin Scalia: Although we have very little information on how Justice Scalia
would vote in an abortion case, his extremely conservative rulings in matters such
as law enforcement and capital punishment led us to believe that he would vote
accordingly on abortion. For this reason, and because he was appointed by Pres-
ident Reagan whose views on abortion are well-known, we gave him a high
ranking on POLITICAL ISSUES (0.230). Also, because he has been critical of earlier
rulings, namely Roe v. Wade, we ranked PRECEDENT low (0.089) and WOMEN’S ISSUES
low (0.058). For BIOLOGICAL, we ranked it low (0.058) because we had little expert
information. Finally, we ranked MORAL ISSUES high (0.565) because of his con-
curring viewpoint in Webster v. Reproductive Health Services: He never directly
ruled on the issue of abortion, but made clear that he did not believe in the
constitutional right of a woman to have an abortion. This led us to conclude that he
believed in fetal rights and, therefore, moral issues should be ranked high.
Clarence Thomas: Because Justice Thomas was just recently appointed, we had
no clear abortion rulings to use in our analysis. However, we do know that he has
ruled very closely with Justice Scalia in almost all cases. In the Wall Street Journal
it is stated that ‘‘…Justice Thomas continues to solidify his alliance with Justice
Scalia. Not only do they nearly always vote on the same side of a case, but Justice
Thomas’s reasoning distinctly resembles that of Justice Scalia’’ [3, p. B6]. for this
reason, we ranked each criterion identical to the rankings for Scalia: WOMEN’S
ISSUES (0.058), PRECEDENT (0.089), MORAL (0.565), POLITICAL (0.230) and BIOLOGICAL
(0.058).
John Paul Stevens: From our research, we found Stevens to hold individual rights
and previous cases in high regard. Therefore, we gave PRECEDENT a high ranking
(0.455). In a ruling in 1980, concerning state funding of abortions Stevens dis-
sented with the majority seeing a ‘‘clear conflict with the decision in Roe v. Wade,
which gave women the right to choose not to bear children. Here, government
286 19 Abortion and the States
stacks the deck against this choice, at least for indigent women’’ [11, p. 160]. Also,
this ruling established his beliefs in women’s individual freedom to choose, and, as
a result, we also ranked WOMEN’S ISSUES high (0.220). Because these beliefs directly
conflict with our definition of MORAL ISSUES we ranked it low (0.054). We also
ranked political very low (0.049) since he is not considered to be a conservative
jurist. Finally, we ranked biological high (0.223) because of his concurring with
the trimester system of determining when, and if, women can have an abortion.
Sandra Day O’Connor: Although, Justice O’Connor has been on the court for
many years, it is unclear how she would rule in an abortion case. She has ruled in a
few cases involving abortion but has not clearly stated her opinion. We know,
however, that Justice O’Connor does not believe abortion to be a fundamental
right deserving ‘‘strict scrutiny.’’ Instead she uses the ‘‘undue burden’’ standard
saying that laws should not burden ‘‘unduly’’ a woman’s right to choose an
abortion. Therefore, our research has led us to believe that Justice O’Connor will
vote to restrict women’s rights by strengthening the rights of the states, but will not
vote to overrule Roe v. Wade.
We ranked the issues that we deemed will weaken Roe high, including
PRECEDENT (0.397) and BIOLOGICAL (0.397) and the other criteria low; including,
WOMEN’S ISSUES (0.110), MORAL (0.048) and POLITICAL (0.048). In his book, Dan
Drucker states [9, p. 164]:
O’Connor, the only woman to sit on the Court, can be expected to play a crucial role in its
deliberations. O’Connor is in the middle; she supports the right to abortion, but she may be
willing to restrict it in some way.
David Souter: Because David Souter was appointed recently, we had very little
expert information to incorporate into our model. We found, however, many
people believing him to be a ‘‘fair and thoughtful jurist’’ who bases his rulings
heavily on prior cases. For this reason, we ranked PRECEDENT extremely high
(0.640). We also found him to have a high regard for individual rights and used
this information to rank WOMEN’S ISSUES high (0.125). We had little information on
his moral beliefs or his feelings related to viability, therefore, we ranked MORAL
ISSUES low (0.069) and BIOLOGICAL ISSUES low (0.068). Finally, although he was
appointed by President Bush and should be politically conservative his disparate
rulings in numerous cases involving issues of individual freedoms caused us to
rank POLITICAL ISSUES fairly low (0.098).
Anthony Kennedy: Anthony Kennedy is also a new justice who has no record
regarding how he will rule on abortion issues. Because he was appointed by
President Reagan and considered extremely conservative by many, we ranked
POLITICAL ISSUES high (0.179). Also, he is known to follow closely prior rulings and,
therefore, we ranked PRECEDENT very high (0.590). For WOMEN’S ISSUES we used a
very low ranking (0.065) based on a statement by Eve Paul, general counsel for
Planned Parenthood which stated [9, p. 164]:
This is a new Court. Kennedy is new and has not previously expressed his views on the
abortion issue.
19.4 Criteria 287
Regarding the remaining criteria, we ranked each quite low because of the lack
of available information. For MORAL ISSUES we gave it a ranking of 0.055 and for
BIOLOGICAL ISSUES we ranked it 0.112.
Harry Blackmun: Justice Blackmun is the only justice that we believe will
definitely rule in favor of upholding a woman’s right to have an abortion. He
concurred in the Roe v. Wade decision and has voted to uphold it in every decision
since. He made it clear that a woman’s right to receive an abortion should not be
weakened in his Webster v. Reproductive Health Services decision [9, p. 180]:
Although today, no less than yesterday, the Constitution and the decisions of this Court
prohibit a state from enacting laws that inhibit women from the meaningful exercise of
that right, a plurality of this Court implicitly invites every state legislature to enact more
and more restrictive abortion regulation in order to provoke more and more test cases, in
the hope that sometime down the line, the court will return the law of procreative freedom
to the severe limitations that generally prevailed in this country before January 22, 1973.
In fact, Justice Blackmun is very clear that the fundamental constitutional right
of women to decide whether to terminate a pregnancy survives. He opined [9,
p. 161]:
I fear for the future. I fear for the liberty and equality of the millions of women who have
lived and come of age in the 16 years since Roe was decided. I fear for the integrity of, and
the public esteem for, this Court.
For obvious reasons, we ranked his criteria high that would uphold Roe v. Wade
in its 1973 form. We ranked WOMEN’S ISSUES high (0.477), PRECEDENT high (0.292)
and BIOLOGICAL ISSUES high (0.125). The criteria that would overturn or severely
weaken Roe v. Wade we ranked low; MORAL ISSUES at 0.0252 and POLITICAL ISSUES at
0.054.
The priorities of the issues with respect to the judges are given in Table 19.1.
Next, we prioritized the alternatives with respect to each judge and each cri-
terion. Table 19.2 summarizes the global priorities we obtained. The priorities of
the alternatives are obtained by adding their weights (Table 19.2) for all the cri-
teria and all the judges. We have: Overturn (0.376), Weaken (0.399) and Uphold
(0.225). It is our belief, based on the AHP model, that the Supreme Court will
288 19 Abortion and the States
uphold at least parts of the Pennsylvania statue and will, as a result, weaken the
rights of women who choose to have an abortion in the state of Pennsylvania.
On June 30 1992, The United States Law Week [13, p. 1,201] reported the
result of the U.S. Supreme decision:
The U.S. Supreme Court yesterday reaffirmed the core principles of Roe v. Wade, 410
U.S. 113 (1973), but at the same time upheld most of a restrictive Pennsylvania abortion
statute. Only a provision requiring a woman to notify her husband before obtaining an
abortion was struck down. (Planned Parenthood of Southeastern Pennsylvania v. Casey,
US SupCt, Nos. 91-744 & 91-902, 6/2/92)
19.5 Conclusions
the people of Pennsylvania, and most likely all citizens of the United States. This
is a very complex problem without a clear cut solution. Certainly the type of
problem suited for analysis by the Analytic Hierarchy Process.
Bibliography
1. Barret PM (1992) Justices and two lower courts face confrontation over abortion, Taxes, The
Wall Str J p B6: C2, 24 Jan 1992
2. Barrett PM (1992) Justices agree to rule on abortion law in Pennsylvania, but May Sidestep
Roe, The Wall Str J p A16: C1, 22 Jan 1992
3. Barrett PM (1992) Thomas emerges as bold new justice with strong dissents in criminal
cases, The Wall Str J p B6: C2, 28 May 1992
4. Brueschke E, Brueschke J (1990) Constitutional law: the future of the abortion controversy
and the role of the supreme court after webster v. reproductive health services. Okla L Rev
43:481–513
5. City of Akron v. (1983) Akron center for reproductive health, Inc. 462 United States Reports 416
6. Clements M (1990) Will they make abortion illegal? What voters say, The Pittsburgh Press,
Parade Magazine, pp 4–6, 17 May 1990
7. Debra G (1991) Press lost a friend on the supreme court, Editor and Publisher 124n1, pp 18
and 20, 5 Jan 1991
8. Doe v. Bolton (1973) 410 Unites States Reports 179
9. Drucker D (1990) Abortion decisions of the supreme court, 1973 through 1989: a
comprehensive review with historical commentary. McFarland and Co., Jefferson
10. Planned Parenthood of Southeastern Pennsylvania v. Casey, US SupCt, Nos. 91-704 and 91-
902, 2 June 1992
11. Rubin ER (1982) Abortion, politics, and the courts. Greenwood Press, Westports
12. Roe v. Wade (1973) 410 United States Reports 113
13. Roe v. Wade (1992) Is reaffirmed but no clear standard emerges, The United States Law
Week 60(51):1201, 30 June 1992
Chapter 20
The Benefits and Costs of Authorizing
Riverboat Gambling
20.1 Introduction
On November 26, 1990, six months after the study on which this chapter is based
was undertaken, the Pennsylvania House of Representatives rejected a bill to
legalize riverboat gambling by a vote of 118-81.
Should riverboat gambling be permitted on Pennsylvania’s rivers and lakes? What
impact would this activity have on the state? What are the potential benefits? What are
the possible costs? These are the questions that many are asking as a bill to authorized
riverboat gambling is being considered by the Pennsylvania House of Representatives.
This chapter aims to determine the importance of the costs and benefits
attributed to riverboat gambling and apply a sophisticated decision-making model
to make a recommendation on the issue. First, the pending legislation is examined
and the positions of those supporting and opposing the bill are delineated. Next,
the Analytical Hierarchy Process (AHP) is applied to the riverboat gambling issue.
Judgments acquired from individuals involved in the decision-making process or
affected by the decision are then described. Finally, the synthesized results are
interpreted, and a recommendation on riverboat gambling legislation is presented.
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 291
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_20, Springer Science+Business Media New York 2012
292 20 The Benefits and Costs of Authorizing Riverboat Gambling
The Excursion Boat Gambling Commission, created under the bill, would
license operators, adopt standards for the gambling operations, and regulate gam-
bling activity. The commission would be funded through licensing and admissions
fees. The initial license fee would be $50,000 with an annual fee of $25 per person-
capacity on each riverboat. The admission fee, to be set by the EBG Commission,
would be obtained for each person embarking on a riverboat gambling excursion. In
addition, the municipality could adopt a local admission fee not to exceed 50 cents.
A wagering tax of 15% would be imposed on adjusted gross receipts from
riverboat gambling. Twenty-five percent of the wagering tax would be distributed
to each county having a home port or port of call, based on the ratio of the number
of passengers embarking from that port to the total number of statewide embar-
kations. Another 25% would be allotted to the municipality, and the remaining
amount would be credited to the General Fund of the state. Additionally, a recent
amendment to the bill requires that $1,000,000 be allocated annually for the
treatment of compulsive behaviors.
Promoters of the bill claim [6], [7], [9] that riverboat gambling will stimulate local
business, encourage economic development, and provide additional tax revenues for
government. They cite examples of riverboat gambling operations in Mississippi,
Iowa, and Illinois which have created new jobs and spawned tourism [8]. Supporters
expect local economies to thrive as riverboat gambling patrons fill hotels, restaurants,
and shops [2]. They also project that the legalized gambling operations will reduce
illegal gambling [4], expand the tax base, and provide tax relief for citizens.
Opponents of the bill believe that legalization legitimizes gambling in the
public mind and thus promotes gambling in both legal and illegal forms [1]. They
expect serious social problems to result from an increase in the number of gam-
blers and compulsive gambling behavior [5]. They fear that riverboat gambling
will lead to increased street crime, prostitution, and drug trafficking along with
corruption, extortion, and bribery among public officials. Opponents claim that
riverboat gambling will promote illegal gambling and attract organized crime to
meet patrons’ demands for credit betting, higher stakes, and avoidance of income
taxes. It is also feared that legalized riverboat gambling will lead to pressures to
allow gambling on Pennsylvania soil [3].
A team of three graduate students from the University of Pittsburgh applied the
Analytical Hierarchy Process (AHP) to analyze Pennsylvania’s pending legislation
on riverboat gambling. A brief definition of the model and an explanation of the
20.2 Problem Analysis 293
An unbiased citizen, who is completing his PhD in Public Affairs, rated increased
social opportunities as the most important benefit and safety/crime issues as the
most important cost impacting his decision on the legislation.
In addition to talking with the relevant decision-makers, the groups which would
be impacted by the legislation were interviewed about the respective issues and
objectives in level 6 of the hierarchies. The following groups were identified as
having a stake in the outcome of the legislation: riverboat operators, citizens of
Pittsburgh, local Pittsburgh businesses, state and local government, other river
users, and environmentalists.
296 20 The Benefits and Costs of Authorizing Riverboat Gambling
The local riverboat operator that was interviewed is strongly in favor of the leg-
islation. The following two benefits are equally important to him: increased rev-
enue and diversification of services. In terms of negative issues, he feels that
increased costs of regulation are extremely more important than competition
because regulation is not something that his operation has any control over.
The citizen of Pittsburgh, who is completing his PhD in Public Affairs, feels that
an increase in the number of jobs in the city is the most important benefit that
could result from the passing of the current legislation. Possible tax relief for the
taxpayers is secondary in importance to this individual. This particular citizen also
thinks that the safety of the Pittsburgh residents may be jeopardized due to the
potential for crime and corruption to increase in the city and surrounding areas.
A representative from a large Pittsburgh hotel believes that the most important
benefit that could result from the legislation is the potential for the number of
tourists being drawn to the area to increase. He could not foresee any costs
associated with riverboat gambling.
The team was unable to obtain the cooperation of a large local gambling operation,
likely to be impacted by this legislation. The assumption has been made that
increased competition for local gambling dollars could result and affect existing
gambling operations in a negative way.
The two barge companies that were interviewed felt that any increase in the
number of boats to be used for riverboat gambling would have no effect on their
current operations in terms of increased traffic on the Pittsburgh waterways.
20.2.2.7 Environmentalist
Damage to environment Potential/selective economic loss Social problems Regulation difficulties and costs
Riverboat operators 0.034 0.038 0.049 0.286
Citizens of Pittsburgh 0.305 0.496 0.405 0.132
Local business 0.030 0.038 0.049 0.042
Government 0.156 0.302 0.405 0.478
Other river users 0.100 0.056 0.047 0.032
Environmentalists 0.375 0.070 0.047 0.031
Riverboat operators Citizens of Pittsburgh Local business Government Other river users Environmentalists Composite priorities
Regulation of gambling 0.5 0.054
Competition for gambling operations 0.5 0.054
Safety and crime 0.637 0.220
Increase traffic and crowds 0.105 0.036
Moral addiction 0.258 0.089
Cannibalism of other legal gambling 1 0.049
Regulation of gambling operations 0.600 0.227
Cannibalism of PA state lottery 0.200 0.076
Gov. law enforc./waste services 0.200 0.076
Inc. river traffic need for safety 1 0.053
Increase pollution 1 0.068
299
300 20 The Benefits and Costs of Authorizing Riverboat Gambling
gambling was considered most costly under moral addiction to gambling, government
regulation, and government services.
The weighted judgments were synthesized to quantify the importance of the hierar-
chical elements (Table 20.3). This process produced a benefit/cost ratio of 0.851/
0.877 for authorizing riverboat gambling and a ratio of 0.149/0.123 for not autho-
rizing riverboat gambling. These results indicate that riverboat gambling should not
be authorized in Allegheny County. In evaluating the sensitivity of the priorities, it
should be noted that a reduction in the importance of the two highest weighted cost
issues would still yield the same alternative. Furthermore, an increase in the impor-
tance of the three highest weighted benefit objectives does not alter the decision.
In evaluating the validity of this AHP application, limitations must be recog-
nized. The biases of the individuals interviewed and of the team members may
slant the judgments thus affecting the results. Interdependencies or commonalities
among elements in the hierarchy possibly existed but were minimized by con-
structing two separate hierarchies. Additionally, imperfect preferences may lead to
inconsistencies among the elements in the model. However, a certain degree of
inconsistency in making judgments is acceptable in real world applications.
USA Today reported January 7, 2010 that ‘‘Pennsylvania legalized poker,
blackjack and other table games at slots casinos Thursday, upping the ante in the
increasingly fierce competition among states for gamblers’ money. Gov. Ed
Rendell, whose signature was the last step in the protracted process of passing the
law, said he had misgivings about expanded gambling, partly because not all of the
14 casinos authorized by the 2004 law that legalized slot machine gambling are up
and running.’’ It took almost ten years since it was proposed.
Bibliography 301
Bibliography
1. Clymer PI (1991) Making waves: riverboat gambling would prompt other betting. Harrisburg
Patriot News, 13 May 1991
2. Guo D (1990) Master salesman puts a new spin on river business. Pittsburgh Post-Gazette, 28
May 1990
3. Legalization of Gambling in Pennsylvania: A Study (1984) League of women voters of
Pennsylvania
4. Pagano C (1987) Superintendent of New Jersey state police, Impact of legalized gambling on
illegal gambling in New Jersey, 16 Oct 1987
5. Schroeder M (1990) Talk About a Riverboat Gambler, Business Week, 72, 23 July 1990
6. Shehan A (1990) Gateway fleet charts cautious course on gambling, Pittsburgh Post-Gazette,
C5, 22 Nov 1990
7. Sheehan A (1990) Riverboat gambling seen bonanza, Pittsburgh Post-Gazette, 4, 14 May 1990
8. Siler C (1990) Why are Las Vegas and Atlantic City yawning? Forbes, 140, 30 April 1990
9. Wolf D (1991) Riverboat gambling bill founders, The Pittsburgh Press, B3, 29 May 1991
Chapter 21
To Drill or Not to Drill: A Synthesis
of Expert Judgments
21.1 Introduction
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 303
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_21, Ó Springer Science+Business Media New York 2012
304 21 To Drill or Not to Drill: A Synthesis of Expert Judgments
The pre-condition for petroleum accumulation is the existence of a source rock, where
oil and gas originate. Once formed, the oil and gas must migrate from the source rock
into more porous and permeable rock called reservoir rock. The reservoir rock acts as
a container for the fluids. A reservoir rock must have enough room to store a sig-
nificant volume of hydrocarbons and must discharge oil or gas readily when the
reservoir is penetrated by a well. In order to accumulate oil or gas, the reservoir rock
must be deformed either by folding or faulting to form structural or stratigraphic traps.
In the structural type, the traps are the result of movements of the Earth’s crust:
folding, faulting, fracturing, or intrusion of a salt dome. A stratigraphic trap occurs
when a porous rock layer is tilted and eroded. The eroded end is then sealed off by
a tight rock layer. Also, to get the petroleum out, there must be some natural
driving force within the reservoir, usually gas and/or water.
Figure 21.1 illustrates the interrelationships of the key factors determining the
volume of recoverable oil in a reservoir and Fig. 21.2 is the hierarchy for estimating
effective porosity (u), oil saturation (So), effective permeability, recovery factor (RF),
and bottom hole pressure (BHP). A brief explanation of each factor is given below.
Table 21.3 Probability distribution for the type of trap and oil saturation
Oil saturation Trap Composite Midrange (S) W9S
weights (W)
T1 (0.167) T2 (0.833)
40–50% S1 0.124 0.129 0.128 0.45 5.76
50–60% S2 0.310 0.l53 0.479 0.55 26.345
60–70% S3 0.469 0.261 0.296 0.65 19.24
70–80% S4 0.064 0.063 0.063 0.75 4.725
80–90% S5 0.032 0.033 0.033 0.85 2.805
EV = 58.875
where u is the midpoint of each range of porosity and P(u) is the probability of
each range represented by W in Table 21.1. According to this table,
EV(u) = 0.105 or 10.5%.
b. Expected value of bottom-hole pressure (BHP). A similar procedure is used to
determine the expected value of the bottom-hole pressure as shown in Table 21.2.
c. Expected value of oil saturation (So). Figure 21.2 shows that oil saturation
range depends on the type of trap. Thus, to compute the expected value of oil
saturation we must first determine the relative likelihood of each kind of trap,
structure and stratigraphic and probability distribution of oil saturation for each
type of trap. We then compute the overall probabilities of the saturation ranges
by multiplying the probability distributions of the oil saturation by the proba-
bility of their corresponding trap as shown in Table 21.3.
d. Expected value of effect permeability (K). Relative permeability depends on
effective porosity (u), bottom-hole pressure (BHP), and oil saturation (So).
Thus, to compute the expected value of the relative permeability for each range
of values for the above variables, we compute the composite weights of each
range of relative permeability as described in [5] (see Tables 21.4, 21.5, 21.6,
21.7).
e. Expected value of recovery factor (RF). Recovery factor depends on oil satu-
ration (So), relative permeability (K) and bottom-hole pressure (BHP). The
computation process for determining the expected value of the recovery factor
is similar to (d) (see Tables 21.8, 21.9 21.10, 21.11).
308 21 To Drill or Not to Drill: A Synthesis of Expert Judgments
Table 21.4 Probability distribution for effective porosity and relative permeability
Effective porosity
Relative u1 u2 u3 u4 u5 Composite
permeability 0.043 0.521 0.238 0.156 0.043 weights (W)
K1 0.513 0.328 0.295 0.059 0.031 0.274
K2 0.261 0.485 0.526 0.305 0.095 0.441
K3 0.129 0.104 0.099 0.430 0.226 0.160
K4 0.063 0.054 0.051 0.147 0.507 0.088
K5 0.033 0.030 0.029 0.059 0.140 0.039
Table 21.5 Probability distribution for bottom-hole pressure and relative permeability
Bottom-hole pressure
Relative permeability L M H Composite weights (W)
0.226 0.674 0.101
K1 0.513 0.102 0.029 0.188
K2 0.261 0.504 0.102 0.409
K3 0.129 0.245 0.287 0.223
K4 0.063 0.102 0.428 0.126
K5 0.033 0.046 0.153 0.054
Table 21.6 Probability distribution for oil saturation and relative permeability
Oil Saturation
Relative S1 S2 S3 S4 S5 Composite
permeability 0.128 0.479 0.296 0.063 0.033 weights
K1 0.513 0.081 0.060 0.034 0.030 0.125
K2 0.261 0.428 0.257 0.201 0.183 0.333
K3 0.129 0.275 0.488 0.420 0.263 0.328
K4 0.063 0.176 0.149 0.270 0.432 0.168
K5 0.033 0.041 0.047 0.076 0.092 0.046
EV = 175.47
Table 21.8 Probability distribution for oil saturation and recovery factor
Oil saturation
Recovery S1 S2 S3 S4 S5 Composite
factor 0.128 0.479 0.296 0.063 0.033 weights
L 0.785 0.731 0.333 0.218 0.167 0.568
M 0.149 0.188 0.592 0.691 0.740 0.352
H 0.066 0.081 0.075 0.091 0.094 0.078
f. Volume of primary recoverable oil. The volume of primary recoverable oil from
one zone can be determined using the following equation:
Bxhx/x So xAxRF
VðSTBÞ ¼
5xD
1:05 þ 100;000
310 21 To Drill or Not to Drill: A Synthesis of Expert Judgments
where
V(STB) is the volume of recoverable oil reserves in stock tank barrels (STB),
B is a constant equal to 7,758 when the volume is measured in barrels and
43,560 when it is measured in cubic feet,
h is the thickness of the reservoir rock in feet,
u is the porosity of the reservoir rock,
So is the oil saturation,
A is the drainage area in acres based on well spacing of 40, 80, 160, or even
640 acres,
RF is the recovery factor, and
D is the depth of the reservoir rock in feet
For the field under study we have:
h = 10 ft
u = 10.5% (from Table 21.1)
So = 59% (from Table 21.3)
A = 40 acres
RF = 22% (from Table 21.11)
D = 2,750 ft
Thus, we obtain:
ð7; 758Þð10Þð0:105Þð0:59Þð40Þð0:22Þ
VðSTBÞ ¼ ¼ 35; 615 barrels:
1:05 þ ð5Þð2750Þ
100;000
Considering the large number of variables in the model and uncertainties associ-
ated with these variables, this is a good estimate. Based on the current rate of
production from a well, the actual volume of primary recoverable oil from Aux
Vases formation is estimated to be between 40,000 and 50,000 barrels of oil.
21.4 Conclusions
Bibliography
22.1 Introduction
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 313
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_22, Ó Springer Science+Business Media New York 2012
314 22 Modeling the Graduate Business School Admissions Process
In order to provide the reader with some background concerning graduate business
admissions, we describe the current selection process that is in place at the Joseph
M. Katz Graduate School of Business, University of Pittsburgh [7].
The director of admissions at the Katz School is responsible for the recruitment,
evaluation, and selection of applicants. The director and her staff first collect and
organize all relevant application materials (such as transcripts and recommenda-
tions) from each candidate. The amount of paper work is substantial. In recent
years, over 3,000 individuals have applied to the school per year. Once a candi-
date’s information file is complete, the admissions staff evaluates and classifies the
application into one of six categories: Automatic accept, Accept, Marginal
accept—accept on recommendation of director, Marginal reject—reject on
recommendation of director, Reject, or Conditional accept—pending satisfaction
of a condition (such as completion of a mathematics requirement). To facilitate the
discussion and selection process, this classification is done prior to submitting the
files to the admissions committee.
In past years, this broad classification was accomplished by comparing an
applicant’s undergraduate GPA and GMAT score to scales that had been
developed by admissions staff members. To preserve the confidentiality of the
school’s admission process, these scales will not be given in detail. For
example, we would automatically accept a student with a GMAT score above
625 and with GPA above 3.30 with other supporting qualifications. Each scale
was divided into categories indicating levels of acceptability which helps in
identifying marginal candidates or candidates who do not qualify. For those
candidates who would qualify or fall in a marginal category, other criteria (such
as letters of recommendation, community services, activities, essays, and work
experience) would be examined further in order to ascertain if there is sufficient
cause to continue the evaluation process. In addition, an interview may be
requested of a candidate to obtain more information or to provide a clarification
of the application materials. After all application information has been reviewed
and the initial classification has been made, the application file is then sub-
mitted to the admissions committee for final review and decision.
22.3 Admissions Selection Model 315
(a)
In the current selection process, we see that, for a single applicant, the admissions
director and her staff and the admissions committee itself are confronted with a large
number of tradeoffs over a diverse set of criteria. When there are several thousand
316 22 Modeling the Graduate Business School Admissions Process
(b)
(c)
(d)
hierarchy. The Major criterion is used to evaluate applicants on the basis of their
previous course of study. It was broken down into broad rating categories: Business
and related majors (such as economics), Engineering, Liberal arts (for example,
English, history, or political science), and Sciences (for example, chemistry,
biology, or nursing). GPA is the actual average from all previous college-level
work that is shown on an applicant’s official transcript. The ratings for GPA are
divided into five categories that range from Over 3.6–Under 2.75 points.
The other level three criterion under Quantitative is GMAT scores. This cri-
terion is subdivided into three scores: Verbal, Math, and Total. The ETS scores the
verbal and mathematical parts of the GMAT are evaluated on a scale from 0 to 60
(scores below 10 or above 46 are rare). An applicant’s total score on the GMAT is
not calculated by summing the verbal and mathematics scores. Rather, the total
score lies between 200 and 800. To account for this, we have subdivided the
Verbal and Math criteria into four ratings that range from 30 and Over to Below
20. The Total criterion is broken down into five ratings that range from 600 and Up
to Under 480.
Also at the third level, the Qualitative criterion is divided into two subcriteria:
Application Information and Work Experience. Application Information covers all
of the data submitted by an applicant and is broken down into four subcriteria:
Essays, Letters of Recommendation, Academic Awards, and Activities. Applicants
are required to submit two essays and can also submit an optional third essay.
These essays are read by the admissions office staff and critiqued for content,
grammar, style, and general writing ability. The entire set of essays are then rated
as Outstanding, Very good, Average, Below average, and Poor.
The second subcriterion under Application Information is Letters of Recom-
mendation. Applicants are required to submit two recommendation forms that
have been completed by employers, professors, or other non-related superiors. On
these forms, respondents rank applicants on several questions and answer several
open-ended questions. The admissions staff then rates the letters as Outstanding,
Very good, Average, Below average, and Poor.
The third subcriterion under Application Information is Academic Awards.
This takes into account an applicant’s university awards, fellowships, membership
in honor societies, and other forms of recognition for academic excellence.
The overall number of awards and prestige of the awards are determined by the
admissions staff and a rating from Outstanding to Poor is then assigned to
the applicant. The final subcriterion, namely Activities, is handled in a similar
way. An applicant’s community and extracurricular activities are rated from
Outstanding to Poor.
The Work Experience subcriterion is broken down into years of Experience and
Type of Work Experience. Years of Experience is a straightforward category with
applicants rated as having experience of 5 years or More, 3–5 years, 1–2 years,
and Less Than 1 year. The Type of Work Experience subcriterion classifies
applicants into four work categories: (1) a business-related job in which the
applicant is considered a professional or management level employee (for exam-
ple, staff accountant, credit analyst, bank manager); (2) a business-related job in
22.3 Admissions Selection Model 319
which the applicant is an hourly or clerical-level worker (for example, bank teller,
secretary, receptionist); (3) a non-business job in which the candidate is considered
a professional (for example, doctor, nurse, dentist);and (4) a non-business job in
which the applicant is not considered a professional or manager (for example,
janitor, construction worker, waitress).
We envision that our AHP model would be used in the following way. At the
beginning of each application period, the admissions committee would meet as a
group to review the overall appropriateness of the model. The model would be
updated: new criteria might be added or old criteria might be deleted in accordance
with the ‘‘qualities’’ that committee members are looking for in a prospective
student. The group would then judge the importance of criteria and subcriteria for
that particular year. In our experience, using the Expert Choice computerized
package [2] greatly facilitates this process. Next, the committee would compare
the performance ratings at the bottom level of the hierarchy. For example, the
relative importance of intensities such as Outstanding, Very good, Average, Below
average, and poor for each criterion would be determined via the expert judgment
of the committee.
To demonstrate this process, the current director of admissions at the Katz
School of Business used Expert Choice to generate the weights shown in each box
of Fig. 22.1 (Of course, the actual judgments necessary to produce the weights
would be made by the admissions committee). From this figure, we see that the
Quantitative criterion with a weight of 0.60 is slightly favored over the Qualitative
criterion with a weight of 0.40.
Once the overall weights have been generated, individual applicants are then
rated. We recommend using a spreadsheet with table lookup functions to carry out
this part of the process. We illustrate this in Table 22.1 with a hypothetical set of
applicants. From this figure, we see that J. Smith is rated as to her performance in
each of the 12 categories. For example, in the School category, she received a Very
Competitive rating. This contributed a weight of 0.028 to her overall score of
0.442. Smith’s score could then be compared with a predetermined threshold; if
the score fell above this value, J. Smith would be admitted. We point out that, with
our model, the maximum score an applicant could achieve is 1.00 and the mini-
mum score is 0.124.
22.5 Conclusions
Using the AHP, we have constructed a model that can provide admissions com-
mittees with help in making complex admissions decisions. We believe that our
modeling process ensures that committee members are in general agreement on
320
up
(1.00) (0.044) (0.101) (0.230) (0.041) (0.052) (0.131) (0.117) (0.023) (0.050) (0.050) (0.096) (0.064)
J. Smith Very Business 2.75–2.9 30 and 30 and 600 and Below Average Below Average Less than Business
competitive over over up average Average 1 year Professional
(0.442) (0.028) (0.016) (0.037) (0.041) (0.052) (0.131) (0.014) (0.009) (.012) (0.019) (0.019) (0.064)
F. Davis Very Lib’l Over 24–26 below 20 510–549 Very Very Average Very Less than Business
competitive art 3.6 good Good Good 1 year non-
professional
(0.557) (0.028) (0.058) (0.230) (0.024) (0.011) (0.050) (0.060) (0.014) (0.019) (0.031) (0.019) (0.013)
C. Jones Comp. t.Eng. 3.0–3.24 below below under Very Average Below Poor 5 or more Business non-
20 20 480 good Average years professional
(0.417) (0.018) (0.101) (0.063) (0.008) (0.011) (0.019) (0.060) (0.009) (0.012) (0.007) (0.096) (0.013)
K. Baker Below average Sciences 3.0–3.24 24–29 30 and 550–559 Average Below Average Poor Below 1 to 2 years Business
over Average Professional
(0.471) (0.003) (0.097) (0.063) (0.024) (0.052) (0.082) (0.029) (0.005) (0.007) (0.012) (0.033) (0.064)
Low score Below average Business Under below 20 below 20 under Poor Poor Poor Poor Less than 1 Other non-
2.75 480 year professional
(0.124) (0.003) (0.016) (0.018) (0.008) (0.011) (0.019) (0.007) (0.003) (0.007) (0.007) (0.019) (0.006)
Remark—Earlier applications of the absolute mode of Expert Choice simply weighted subcriteria weights by criteria weights, then intensities by subcriteria weights to obtain the overall weights of the
intensities. In this model we have modified the former approach. To produce a ratio scale that is part of the unit priority of the top level goal would be to divide the intensities with respect to each criterion by
the maximum intensity under it thus producing a unit for the intensities with respect to each criterion. The overall rating of the alternatives would then belong to a ratio scale which when normalized
becomes a proportionate allocation of the unit weight of the goal to each alternative
Modeling the Graduate Business School Admissions Process
22.5 Conclusions 321
admission goals. Our approach is systematic and thorough: it effectively uses all
information supplied by an applicant. The most obvious advantage of using our
model for admissions selection is that it provides for consistent decision making.
All applicants are evaluated on a single set of weighted criteria. This should help
to reduce the subjectivity of the process. Again, we wish to emphasize that our
model does not replace directors of admissions, staff personnel, and committees.
Rather, it should be used as an aid in the overall decision-making process.
Early in the spring of 1991, we formally presented our model to the graduate
admissions committee of the Katz School. The committee viewed our model
favorably and believes that it is a tool that can be incorporated into the evaluation
and selection process. The committee plans to refine the hierarchy, modify the
judgments, and generate a new set of weights. Through use of the model, the
hierarchy will no doubt be expanded to accommodate other factors such as
graduate education and part-time work experience that need to be considered in the
admissions process.
Bibliography
23.1 Introduction
Infertility is perceived as a major crisis in life. There are strong religious, cultural
and societal pressures to have children. The most commonly accepted definition of
infertility is the continued inability of a couple to conceive after a year. The high
level of interest in the problems associated with infertility has led to the rapid
development of medical technology in this area. It is one of the few specialty fields
that have given rise to many emotional, ethical and legal considerations.
It is estimated that 11% of all married couples with the woman between the
ages of 15 and 44 had problems with conception. An increasing number of infertile
couples are seeking help. It has become more socially acceptable to seek fertility
care. The number of US infertility clinics performing high-technology procedures
increased from 84 in 1985 to 270 in 1992. The procedures are quite costly.
Infertility has become an industry with revenues of $2 billion a year [4].
Confronted with the reality of infertility, couples have several options: remain
childless, adoption, undergo in vitro fertilization (IVF) or artificial insemination,
or become parents with the cooperation of a surrogate mother. The effects of
childlessness and its alternatives on peoples’ quality of life and on their
self-perceptions provide many challenges for modern mental health professionals.
Not only do we need to know what alternatives exist and their potential benefits,
but also the costs and risks involved with each. Related psychological and legal
issues have increasingly come to the fore. The feelings of frustration, self-doubt
and negative self image are real and as painful as the medical procedures them-
selves. In addition to these considerations, there is another major factor—the
financial cost inherent in adoption, biotechnological procedures and surrogacy.
The extensive medical efforts vary from hundreds to thousands of dollars per
procedure or attempt. Adoption agencies as well as ‘‘brokers’’ charge for their
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 323
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_23, Ó Springer Science+Business Media New York 2012
324 23 Infertility Decision Making
services. Surrogate mothers frequently assume that status from around $10,000 to
$15,000. Attorneys charge anywhere from their normal hourly rate for the paper
work to $10,000 or more for their ‘‘matchmaking’’ role.
The complex problem of arriving at the best option for having a baby when it is not
possible by natural process is analyzed by the Analytical Hierarchy Process
(AHP). The AHP is a process of solving complex problems by logic and systemic
rationality. It incorporates both the qualitative and quantitative aspects of human
thought in the decision making process. The problem is defined and the hierarchy
constructed in a qualitative manner and the judgments made qualitatively.
It enables one to study the problem as a whole while taking into consideration the
interactions between the components within the hierarchy. Six alternatives were
chosen for consideration as a possible solution to the problem. A description of
each alternative follows.
23.2.1 Adoption
Adoption is defined as ‘‘the overall legal process by which a parent who is not the
natural parent of a child becomes legally recognized as that child’s parent’’. It is
often the last resort of infertile couples in trying to establish a family. Many of the
couples have pursued medical therapies and most have rejected the legally and
emotionally risky option of surrogate motherhood. In 1986 it was estimated that
there were 60,000 nonrelative adoptions [6]. Until the 1970s, the most common form
of nonrelative child placement was through public agencies. The most common
criticism of agency adoption was that prospective parents had to sometimes wait for
as long as 7 years to receive a child placement. Private or independent agencies were
established in the following years. They charged ten times the fees charged by public
agencies, $10,000 compared to $1,000 by the public agencies. However, they
became successful as the wait period was only 12–18 months and couples were
willing to pay the price for that. This form of adoption still has an intermediary, but it
is more likely to be an attorney or physician rather than a state agency. The legal
documents are still drawn in court and the judge must grant the adoption.
A number of court cases have considered the constitutionality of racial
matching in adoption. The standard commonly used in making an assessment of
whether a particular home would benefit the child is that the placement be in the
‘‘best interest of the child’’. Several federal courts have reviewed statutes that
make racial matching a factor in adoption. In all of these rulings, the courts have
held that race may be a factor, if it is in the best interest of the child, but it cannot
be the automatic factor in determining placement of the child.
23.2 The Alternatives 325
The changes in adoption practice and laws have made this a somewhat more
hazardous alternative to childlessness than was true one or two generations ago.
Yet there is no question that the majority of adoptions go through the courts
without problems and result in happy families. Today, more than in previous years,
there may be a need for mental health counseling and effective legal advice at the
initial stage when adoption is considered and then throughout the adoption
process.
In vitro fertilization (IVF), also known as the creation of ‘‘test tube babies’’, has
been performed since the late 1970s. It involves extracting ova from the woman,
fertilizing with the husband’s sperm in a glass petri dish, and then after several
days of growth reimplanting it in the woman’s womb. One of the striking features
of this procedure is the wide variation of success rates due to the complex nature of
the technique and consequently the numerous variables which may impact the
success. Some clinics have established a higher success rate and couples gravitate
toward them. Even then the success rate is at most only 12 or 13%. It is a rather
expensive procedure, costing between $6,000 and $10,000 per cycle including
medication. The high cost is due to tight quality control and individual care which
produce good results [7]. IVF has spawned a burgeoning industry that is almost
326 23 Infertility Decision Making
devoid of regulation. As a result, over 200 clinics have sprung up across the
country. Yet, approximately 1/3 of these clinics have not produced successful
births. The current IVF market is estimated at $30–$40 million per year [3].
At present there are no specific federal statutes concerning assisted reproduc-
tion. Two bills may be introduced in the near future in Congress that could have an
effect of the practice of IVF. The greatest controversy concerns the legal standing
of embryos. The two sides of the controversy are that the embryos are property
and, conversely, that they are humans.
23.2.4 Surrogacy
23.3 Procedure
The Analytical Hierarchy Process was applied to benefit/cost analysis to obtain the
overall best solution for the problem. The benefit (success) of each alternative was
weighed against the cost and risk involved and a ratio of benefit/cost 9 risk was
calculated. Three separate hierarchies were constructed for the three parameters,
viz., benefit, cost and risk. The results were then integrated to obtain the best
solution. The three hierarchies are discussed below:
In order to achieve maximum benefits with the alternatives, the criteria considered
most important were (1) genetic identity, (2) legality, and (3) gender selection
(Fig. 23.1).
Genetic Identity: With advances in biomedical technology it is now possible to
screen a potential mother or father for congenital defects. The probabilities of
passing on a trait to an offspring can be calculated and in many instances
23.3 Procedure 327
23.3.2 Costs
There are three types of costs that were taken into consideration in determining the
best alternative: (1) monetary, (2) physical, and (3) emotional costs. Monetary cost
is tangible and physical and emotional costs are intangible costs. One of the
characteristics of AHP is that it allows one to measure such intangible costs.
Monetary cost: In addition to considerable physical and psychological costs and
time away from work, infertility treatment is expensive [2, 5]. According to Office
of Technology Assessment estimates [1], the total cost to treat infertile couples
ranges from $3,660 for the simplest problem to $22,200 for the whole treatment if
the couple goes all the way through IVF. Not all insurance companies will cover
328 23 Infertility Decision Making
infertility as part of the benefit package; therefore, monetary cost may affect the
patient’s choice of treatment. In evaluating the costs involved in the entire process,
three factors contributed to the expenses. These constitute the subcriteria in the
hierarchy. The first expense a couple encounter is the cost of evaluating the various
alternatives. These are the fees paid for medical screening in IVF and artificial
insemination procedures and initial consultation fees paid to attorney or adoption
agency for surrogacy and adoption (Fig. 23.2).
The next expense is the cost of the actual procedure. This includes expenses
incurred with the surgical procedure in IVF and artificial insemination. Payment to
the surrogate mother and her prenatal care and childbirth are incurred in the case of
surrogacy. Fees paid to the adoption agency may be considered as the adoption
expense.
The third expense is the legal fee paid to the attorney. Adoption requires a legal
contract to be drawn. It also requires filing a petition with the court and obtaining a
ruling. Laws allowing surrogacy vary from state to state. Of the states allowing
surrogacy, no state has set a detailed regulatory scheme to establish the legal rights
and obligation between the parties and the child. If allowed, a surrogacy
arrangement would require a legal contract drawn by an attorney and court
approval prior to initiation of a pregnancy.
Physical cost: Fertility rate in women is highly dependent on age. Late teen and
early twenties is the most fertile period in a woman’s life. Infertility increases with
age and significantly so after the age of 35. The infertility rate for women between
the ages of 35 and 39 is 24.6% and for women between the ages of 40 and 44 is
27.2% [8]. Hence success in medical procedures to overcome infertility is highly
dependent on the age of the person carrying the baby. In the case of surrogacy
where the biological mother’s egg is used, the age of both the biological mother
and surrogate become important. Another physical factor equally important is the
state of health of the person carrying the baby.
Emotional cost: Infertility creates psychological distress in itself. In addition,
they are faced with a confusing array of options, as well as uncharted ethical and
legal implications that are very stressful. Support groups have proved to be
valuable. Infertility makes couples more vulnerable to depression. The emotional
23.3 Procedure 329
23.3.3 Risk
23.4 Results
The decision reached by the AHP is discussed for each hierarchy followed by
benefit/(cost 9 risk) analysis. Note that risk is in the denominator multiplied by
the cost because one asks: which person is more at risk and which alternative has
greater risk? The weighted judgments for each hierarchy were synthesized. Of the
three criteria for the benefits of the process, genetic identity was the most
important. IVF was found to be the most beneficial solution, but at the same time
the riskiest procedure and also one of the most expensive. Surrogate mother using
biological egg and sperm was found to be the least beneficial and at the same time
most expensive. Artificial insemination was the most cost effective.
330 23 Infertility Decision Making
Benefits were weighed against cost and risk taken together. A benefit/cost ratio
by AHP showed AIH as the best solution for a couple who are unable to have a
baby naturally (Table 23.1).
In evaluating the validity of AHP in the problem, the limitations of the model
must be recognized. The results depend upon the criteria chosen to evaluate the
benefits and costs and also the weights given to the judgments by the individual or
individuals conducting the study. However, consistency is checked to maintain
coherence among the judgments. While perfect consistency is not always possible,
and often not even desirable, inconsistency of up to 10% is tolerated. If incon-
sistency exceeds 10%, the AHP model allows one to retrace the steps and revisit
the judgments.
Given the complexity of the issues in this problem and the rapid technological
advances in this area, the decision may change in the future. However, under the
present circumstances, we feel that we have considered all the important and
relevant criteria in the study to offer the best solution to the problem.
Bibliography
1. Congress of the United States (1988) Office of technology assessment. Infertility: medical and
social choices. Washington DC: US Government Printing Office (OTA-BA-358)
2. Cooper GS (1986) An analysis of the costs of infertility treatment. Am J Public Health 76:1018
3. Cosco J (1988) Baby boom. Public Relat J 44(2):28
4. Dewitt PM (1993) In pursuit of pregnancy. Am Demographics 15(5):48
5. Fuchs VR, Perrault L (1986) Expenditures for reproduction related health care. JAMA 255:76
6. Gibbs N (1989) The baby chase. Time 9:86
7. Meldrum DR (1993) In vitro fertilization. In: Schlaff WD, Rock J (eds) A decision making in
reproductive endocrinology. Blackwell Scientific Publication, Boston
8. Pratt WF, Mosher WD, Bachrach CA, Horn MC (1984) Understanding US fertility: findings
from the national survey of family growth, cycle III. Populat Bull 39(5):1
9. Yeh J, Yeh MU (1991) Legal aspects of infertility. Blackwell Scientific Publications, Boston
Chapter 24
Deciding Between Angioplasty
and Coronary Artery Bypass Surgery
24.1 Problem
This chapter is an effort to create a decision process for patients and physicians to
decide between open-heart bypass surgery or the less intrusive angioplasty
procedure. The model was developed with ‘‘Expert Choice for Windows vs. 9.0,’’
including pairwise comparisons of benefits, of risks and of both procedures. The
overall result slightly favors angioplasty because it indicates a lower risk potential
and in addition seems to be more beneficial for the patient. The results that we
determined with the AHP model, support some of the research in the medical
community, but not all evidence favors angioplasty.
More than two decades ago, medical scientists discovered a new procedure to
treat heart disease (see Fig. 24.1) besides using the conventional bypass surgery.
The bypass technique required a vein from elsewhere in the patients’ body be
taken and used to bypass the blocked artery (See Fig. 24.2).
Angioplasty is a newer procedure in which a balloon is expanded in a narrowed
blood vessel, stretching the walls of the vessel to restore normal blood flow. In 1996,
666,000 (452,000 men and 214,000 women) angioplasties were performed in the
U.S. Angioplasty is mainly used if a patient has coronary artery disease (CAD),
which is the most common form of heart disease in the U.S. CAD means that blood
vessels that feed the patients’ heart are narrowed or blocked (See Fig. 24.3).
During angioplasty, a catheter (a thin, flexible tube) with a balloon at the tip is
inserted into the patients’ artery to widen the passageway (see Fig. 24.4). Then, the
catheter is removed. After the procedure, the patient may need to stay in the
hospital for a day or more.
If a patient needs treatment for serious CAD, he/she and the doctor face one of
the above-described prospects. Consequently, the central question the physician
seeks to answer is whether to recommend bypass surgery or angioplasty. By using
the AHP, the following discussion examines the various risks and benefits that are
inherent in the two methods.
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 331
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6_24, Springer Science+Business Media New York 2012
332 24 Deciding Between Angioplasty and Coronary Artery
Two models were built to answer the question of whether angioplasty is preferable to
the more traditional bypass operation. The first model weighs the ‘‘benefits’’ of the
angioplasty when it is compared to the bypass; several factors were taken into
consideration and weighed accordingly. The second model did the same but weighed
the most important ‘‘risks’’ of the angioplasty when compared to the bypass proce-
dure. The following sections will detail each level of the model and attempt to explain
its importance to the decision making process and why it is included in the analysis.
Under the ‘‘Benefits’’ model there are two sub-headings, technical and patient
benefits. Each is detailed in the following synopsis. The major benefit of the
angioplasty procedure is that actual surgery is not required and it is less invasive to
the patient. In addition, a single angioplasty procedure is less expensive then a
bypass operation.
Age: For this model, age is the most important factor (see Table 24.1). With
advancing patient age, angioplasty is a more suitable option since its higher
recurrence rate is less likely to matter given shorter life expectancy.
Medical Comorbidities: The following comorbidities are included:
334 24 Deciding Between Angioplasty and Coronary Artery
• Hypertension
• Diabetes
• Renal failure
• Hepatic failure
• Systemic cancer
• High general risk for general anesthesia
• Poor rehabilitation candidate
The presence of any of these conditions would tend to favor angioplasty to
varying degrees. For example, a bypass surgery for diabetics is better in com-
parison to an angioplasty because there is greater tendency for their arteries to
reclog after angioplasty.
Patient Symptoms: Patient symptoms in the model include acute chest pain and
congestive heart failure. Patient symptoms is one of the less important factors in
the model. Patients with ongoing chest pain or congestive heart failure are better
candidates for angioplasty, which can often relieve these symptoms quickly
without the need for prolonged recovery following bypass.
Cardiac History: The most important factors of a patient’s prior cardiac history
in this model are:
• Lack of prior angioplasty and/or stenting
• Presence of prior bypass
• Lack of need for valve repair
Patients who have not had either prior angioplasty or stenting procedures are
better candidates for angioplasty compared to those who have. In addition, patients
who have had, and failed, a prior bypass are increasingly considered better can-
didates for angioplasty. Patients who need an open operation for valve repair
would not be good candidates for angioplasty since they need an open procedure
anyway.
Cardiac Pathology: This factor is rather important in our model (see
Table 24.1). The following three elements are included in this factor:
• single vessel stenosis
• lack of complete occlusions
• less than 50% blockage of the left main coronary artery
Patients with stenosis (narrowing of a vessel) of only a single blood vessel are
the best candidates for angioplasty; this is in contrast to patients who have had
complete blockages of all three major arteries to the heart. If more than 50% of the
left main coronary artery is blocked, it is better to have an open bypass operation.
Patient Preference: A rather less important factor is a patient’s preference. The
patient can have religious reasons for not allowing blood transfusions (e.g.
Jehovah witnesses). In addition, it is necessary to pay attention to a patient’s fear
of general anesthesia and his/her refusal of surgical incision. Patients with a
stronger preference for minimally invasive procedures or inability to have blood
products transfused would be more strongly considered for angioplasty.
24.2 The Model 335
Physician Factors: The physician factors in the model are rather unimportant.
They include experience, skill, and personal preference. Physicians with more
experience and greater skill with angioplasty would tend to recommend angio-
plasty. Experience matters, especially in complex procedures like angioplasty.
Standards recommend that a physician perform at least 75 angioplasties per year
and that an institution perform at least 200 of these procedures annually.
effectively the blockage was eliminated. Also the patient would be put on an
exercise program and his/her doctor would likely to see him/her several times a
year to make sure no more blockages have occurred.
Expense: This analysis does not outline the quantifiable costs explicitly because
these costs are about the same for both procedures. The angioplasty treatment costs
less originally, but clots form so often that it has to be repeated, which increases
the total costs.
Angioplasty blocks the artery briefly when the balloon is inflated and consequently
the artery collapses immediately, which can trigger a heart attack if the blood
supply to the heart muscle is already too severely restricted. The procedure can
also damage or puncture an artery. In addition, the risk of blood clots increases
slightly (as the catheter touches the artery wall). Another disadvantage of the
method is that the artery frequently becomes blocked again, requiring another
angioplasty or surgery.
There are in fact some potential negative side effects, which can occur after the
angioplasty, e.g. the arm or leg (at the insertion site) may feel numb or cold, there
may be signs of infection (redness or oozing) appearing at the insertion site or
there may be chest pain and discomfort.
In addition, angioplasty patients have a higher likelihood of experiencing
angina (chest and arm pain) and exercise-induced ischemia (insufficient supply of
blood to a particular organ or tissue). In order to become angina-free over time,
patients must undergo repeat procedures of revascularization (through either
angioplasty and/or bypass surgery). This in turn increases all the other mentioned
risks connected to both techniques. However, a new method of angioplasty using a
stent can decrease repeated revascularizations.
The preceding factors along with others were considered in the model. As in the
‘‘Benefit’’ model, the ‘‘Risk’’ model also has two sub-headings, patient and technical risks.
The technical risks include five equally weighted factors. However, the patient risks
include eight separate factors with varying weights according to their importance.
Medical Comorbidities: The first heading under the technical risks is comorbidi-
ties, which includes the anethesia risk and the rehabilitation risk for this procedure.
The anethesia risk is weighted much higher because although it is an often
practiced technique, and one can never be sure of a patient’s reaction to it (see
Table 24.2).
Cardiac Pathology: The next technical consideration is the pathology of the
procedure. Under this heading there are three equally weighted factors. These
338 24 Deciding Between Angioplasty and Coronary Artery
factors are: single vessel stenosis, lack of total vessel occlusion and total vessel
blockage.
Physician Factors: This technical consideration is rather arbitrary and depends
on luck of the draw more than anything else. Under this heading there are three
24.2 The Model 339
factors. The most important is the skill of the physician performing the procedure.
This is closely followed by a related factor, the experience of the surgeon. Both are
followed by the personal preference about what he/she believes is the most
effective procedure.
Cardiac History: The cardiac history of the patient to receive the treatment is
one that cannot be forgotten when considering a choice between the two proce-
dures. The risks under cardiac history are equally weighted between three factors.
The first is whether the patient has ever had an angioplasty performed before. The
next factor is similar and that is whether the patient has any prior stenting. The
final historical factor is if the patient has ever undergone a bypass procedure.
Age: The final technical consideration when weighting the risks of the two
procedures is the age of the patient at the time of the procedure. Obviously, due to
its less intrusive nature, angioplasty is preferred for older patients and for the same
reason for younger patients who hope to avoid the bypass procedure.
Death: With any type of medical procedure there is always the chance of death.
This factor is weighted very heavily in the model, because it is a very significant
risk factor for the patient.
Blood Loss: The risk of blood loss is very low in angioplasty, therefore it has
the lowest weight in the model (see Table 24.2). Two types of blood loss are
considered; the most important is loss that would require transfusion.
Infection: Even in today’s pristine operating conditions, there is always a
chance of infection when a surgical procedure is performed. The most common in
these two procedures is the sternal infection that can result from a bypass pro-
cedure. With the angioplasty one associates groin infection where the incision is
made during the procedure.
Lung: Difficulties can arise in a patient’s lung when the bypass procedure is
performed. Pulmonary edema is the problem that has the largest weight in the
model, closely followed by pneumothorax.
Sudden Occlusion: Sometimes, following angioplasty or bypass surgery, one of
the major coronary arteries can become suddenly blocked. This can lead to death
of the cardiac muscle that it supplies, which is more common with angioplasty.
Arrhythmia: This is a condition that occurs when the heart muscle is irritated
and its electrical activity altered. The more common form is atrial arrhythmia
followed by ventricular arrhythmia, with the latter being much more dangerous to
the patient. Atrial arrhythmia is not uncommon after bypass surgery and is less
common after angioplasty. Fortunately, ventricular arrhythmia is rare following
both procedures.
Disease recurrence: There always exists the danger of recurrence of the
symptoms. The model lists the two most common: myocardial infarction and chest
pain. In general, disease recurrence is much more common following angioplasty.
340 24 Deciding Between Angioplasty and Coronary Artery
To test the model, it was appropriate to use it on two actual cases for which the
decision as to whether to proceed with bypass or instead use angioplasty was
known. The first patient was a 52-year-old man with all the symptoms associated
with the onset of heart disease. He is a smoker, asthenic, with high cholesterol and
most importantly a strong family history of coronary artery disease. This patient
had a history of mild chest pain when he exerted himself, such as when shoveling
snow. One day he experienced severe chest pain and subsequently had a heart
attack. Imaging at the hospital indicated severe triple vessel disease with complete
occlusion of two to three vessels. Clinically this patient would be selected for the
bypass procedure because of the high risk of recurrence, the severity of his disease
and because he had complete occlusions. When this patient’s characteristics were
put through the AHP model, it also favored the bypass procedure.
The second clinical case was an 87-year-old woman, who remained an active
non-smoker and had no family history of heart disease. She was a diabetic who
experienced mild renal failure. With no prior history of chest pain, she developed
severe chest pain while walking. Imaging at the hospital indicated 85% stenosis of
the left main coronary artery, the other three vessels showed no damage. Clinically
this patient would be selected for angioplasty due to her low risk of recurrence,
high risk for surgery given her age and diabetes. The AHP model also favored
angioplasty for her.
In general, angioplasty has slightly greater benefits (0.528 versus 0.472) and
slightly lower risks (0.445 versus 0.555) (see Tables 24.1 and 24.2).
24.4 Conclusion
heart condition from getting worse and may even improve the health of his/her
heart.
The outcome of the analysis using the AHP indicates only a very slight ten-
dency to rather do an angioplasty instead of the traditional and more definitive
open-heart bypass surgery. This outcome appears to be a consequence of the great
similarities in the effects of the two surgical techniques (e.g. long-term survival
rates and quality of life measures).
However, the decision whether to do an angioplasty or an open-heart bypass
surgery should consider each patient on an individual basis. Doing either proce-
dure itself must be considered on an individual basis as well.
Index
A C
Abortion, 281 Clustering. See homogeneity, 32
Admissions, 313 Coefficients
model, 314 technological, 107
AHP, 1 Composition
axioms, 4 additive, 25
basic approach, 1 multiplicative, 25, 27
decomposition, 2 Concession
distributive mode, 10 tradeoffs, 233
ideal mode, 10 Conflict
macroeconomic policy, 193 Middle East, 203
marketing strategy, 160 concessions, 212
philosophy, 3 issues, 211
practice, 3 retributive, 206
rationality, 11 Consistency
seven pillars, 24 index, 9
Analysis ratio, 9
benefit/cost, 7 Countries
risk, 7 less developed, 115
Angioplasty, 333 market attractivenes, 129
ANP, 24 Country
Architectural ratings, 137
contiguity, 51
design, 41
needs, 41, 42 D
spaces, 47 Dependence, 107
Assisted suicide, 249 Descriptive, 11
theory, 11
Design
B bridge, 95
Bridge See design marketing strategy, 160
types, 95 sailboat, 83
Budget allocation, 41, 44 Drilling, 303
T. L. Saaty and L. G. Vargas, Models, Methods, Concepts & Applications of the Analytic 343
Hierarchy Process, International Series in Operations Research & Management Science 175,
DOI: 10.1007/978-1-4614-3597-6, Ó Springer Science+Business Media New York 2012
344 Index
N S
Network Scale
composition, 10 fundamental, 3
structure, 3 ratio, 3
Normative, 10 Strategy
theory, 10 pricing, 161
product pricing, 160
Supermatrix. See ANP, 177
P
Petroleum
exploration, 303 T
Priorities Technological. See coefficients
approximation, 8 Technology
vector of, 7 transfer, 120
R W
Rank Weber, 29
change, 10 Weber–Fechner, 29